Mid-2014 county population estimates out tomorrow, March 26

[muon2]

I like the first AZ plan.

The reservations are indeed a problem area. They are like munis that cross county lines, but there's more desire to keep them together than munis at county lines. At a minimum I would agree that there should be no chop penalty for an area that is exactly coincident with the reservation land in a county. At one point jimrtex and I suggested treating reservations as separate counties since that would then create a chop penalty for reservation splits, but not for chopping into the counties to keep them whole. It all depends on how strong one wants to make the incentive to keep reservations whole.

I think I have come up with the right objective function here. I like the preference metric as a way to resolve some of these issues. I will be getting to your King County Pub gerrymander post/map soon. I have quite a bit to say about that one.

Can you write your reservation preference metric in concise words? It seems plausible but I'd like to test it.

On the AZ-3 bridge chop issue, i think it comes down to the fact that we agreed, and I still believe, that there is no preference for CDs nested in a county - it's just about chops. I can rearrange your AZ-3 to put Tucson with Santa Cruz county and the rest of Pima with Yuma and SW Maricopa. The bridge chop vanishes with no extra chops.

[muon2]

I thought that single county UCCs were given no special status. When we considered it in other states it created double penalties of the sort one should avoid.

I think that the bridge chop issue should not be measured with SKEW or POLARIZATION factors. We only consider those after a set of plans has made the Pareto cut.

I know that bridge chops in urban counties can shift the political numbers. They are very good at that. It's for exactly that reason I want to discourage them. If political factors come in after the Pareto cut, then I can use bridge chops to shift CDs a few points in the direction I want, often without any erosity consequences as you have noted. That gets my partisan map through the cut, perhaps at the expense of a neutral plan.

[muon2]

I thought that single county UCCs were given no special status. When we considered it in other states it created double penalties of the sort one should avoid.

I think that the bridge chop issue should not be measured with SKEW or POLARIZATION factors. We only consider those after a set of plans has made the Pareto cut.

I know that bridge chops in urban counties can shift the political numbers. They are very good at that. It's for exactly that reason I want to discourage them. If political factors come in after the Pareto cut, then I can use bridge chops to shift CDs a few points in the direction I want, often without any erosity consequences as you have noted. That gets my partisan map through the cut, perhaps at the expense of a neutral plan.

1. Single county UCC's, or any county having a population in excess of one CD (same thing), in general yes, should not get special status. But with respect to the pack issue, perhaps they should. It is a policy issue.

I think that's a fundamental change in philosophy, and I'm not convinced of the need to change.

You must be logged in to read this quote.

If I can shift CDs by 3% without trying by going from bridge to not bridge, I bet I can do better if I need to. My point is they are very effective at shifting partisan numbers, and often incur no erosity penaly compared to a straight chop. This is an open mapping process and if there's a way to game the system for political advantage, it will be used.

You must be logged in to read this quote.

We've never used the political measures as part of the map filter. The idea has been to find the Pareto frontier from non-political data, then pass those plans along with the political data to the decision making body. They could then use the political scores as they want.

You must be logged in to read this quote.

The SKEW and POLARIZATION will appeal to good government groups, but I am not sure that including them in the selection criteria will sell it to the majority party. They were part of the OH competition and forced some strange choices knowing that they affected selection. Dems liked their impact in OH but not the Pubs.

You must be logged in to read this quote.

I'm quite leery of point 3. If some scores (eg INEQUALITY) are tie breakers, it becomes confusing to have parts of other scores (CHOP) also act as tie breakers. A score should just be a number to make this fly. Perhaps I can suggest a revision of the three points into two as follows:

1. A reservation area within a county is automatically a subunit of the county. Contiguous reservation subunits form a Reservation Subunit Cluster (RSC). The CHOP score shall increase by one for each district over the minimum needed for the cover of the RSC.

2. A chop into a county that only includes a reservation subunit shall not count as a chop of that county if the RSC is covered by the minimum number of districts.

[muon2]

Here's New Hampshire based on population projections. The location of the chop was driven by the equality metric. Even though in my world, equality is at the absolute bottom of the heap, everything else was equal, so that was all that remained as the tie breaker as to the precise location of the chop. Also interesting is that both CD's have almost the same partisan complexion circa 2008, 1.28% and 1.06% Dem PVI's. The Dems would be pleased as NH trends slowly their way (or if you believe one poster here, at WARP speed).  

Were these just from county projections? If you want to get into the weeds the Census has estimates for all 259 towns in NH. When I looked at the muon rules applied to NH it made more sense to ignore counties but look at NECTA's (clusters of towns by urban center) as a replacement for the UCCs.



In any state where one can easily make a chop-free plan (such as with towns in NH or counties in IA) inequality becomes the variable of choice to contrast with erosity for Pareto purposes. Otherwise one has a one-dimensional representation and no flexibility.

[muon2]

I am more interested in skew and polarization than equality. We just disagree on that. And I am getting nervous that you are setting up different rules for different states. You seem searching for more clusters of stuff, be it ethic clusters, and now this variant of clusters. I am not sure what is driving this, other than perhaps boredom or something.

Using town estimates would only make a difference here for the size of the chop, and I am quite sure that this is still the best place to chop for equality purposes. The population number is small, and NH growth rates don't vary that much across the state.

My link is from two years ago and was hardly due to boredom. Counties don't even exist in half of New England as units of government, and they have a minimal role elsewhere. The UCC concept doesn't function well in New England. The Census recognizes the special nature of those six states and defined the New England City and Town Areas accordingly.

You must be logged in to read this quote.

Some states have no predefined county subunits, and the rules have to adapt to that. New England maps much better with towns, not counties, as the basic unit. You'll see a number of examples with 2010 data on the thread I linked on my previous post.

[muon2]

Even if SKEW and POLARIZATION were used to contrast against EROSITY, I don't think it would do much good in states like RI and MA. INEQUALITY is the only game left in town.

[muon2]

Even if SKEW and POLARIZATION were used to contrast against EROSITY, I don't think it would do much good in states like RI and MA. INEQUALITY is the only game left in town.

Yes, of course, and it turned out to be that way in NH, using standardized metrics, rather than custom metrics developed by your creative little mind. RI is losing a seat, so the map there should be really easy to draw in any event.

Mass counties seem to have functions too by the way. I came across that trying to check on the demographics of Lincoln in connection with how rich places vote.

The counties exist a geographic entities, but lack an elected government other than certain officers like sheriff. Here's the description from the the Sec of State.

You must be logged in to read this quote.

The town is vastly more important in MA and the rest of New England than anywhere else. The small scale, few counties, importance of the town, and recognition of such by the Census suggest that towns are the fundamental unit for drawing districts in those states. Based on all of those factors I've been treating those states in terms of towns as I've developed my rules for almost four years now.

[muon2]

This is not a wholly theoretical exercise, because NY now has a redistricting law that will cut way back on the gerrymander regime. So reasonable maps are now the order of the day. This map is reasonable, and also has a good skew of Pub plus 1., and low polarization at 33 (2R, 4r, 5e, 5d, 10D).  Two CD’s are just barely r, and if both went to e, the skew would be zero. Both maps have the same skew and polarization scores, although the first map has NY-17 going from e to r, and NY-22 going from r to e.

I get a different value of SKEW from your distribution of seats. If I assume that NY is still D+11 in 2022, then 26 seats gives a skew offset of 11 D seats (11% * 4 * 26). Your distribution has 15 D+d seats and 6 R+r seats for a difference of 9 D seats. The SKEW is 2 R.

BTW I know you looked into this last spring. Are there significant changes from that plan to his one?

[muon2]

Here is what happens to NY-18 if the NYC metro area for the remaining 5.75 years of the decade grows at a rate of 0.74% per year, rather than 0.69% per year,  ending the incumbent's career in what is now NY-19, assuming NY faithfully follows its new redistricting law. Greenburgh town is chopped, and NY-18 has a Dem PVI of 1.3%, while NY-17 has a Dem PVI of 0.5%, circa 2008. Columbia County really is on the cusp as to where it ends up. If growth slows down  a tad, rather than speeds up a tad, a Muon2 metric map would chop Geeene County, but is what is more likely to happen, is that Ulster County would be chopped, with Columbia County still in NY-18.

In other words, Columbia County goes to NY-22 only in the narrow window of current population growth rates in the NYC metro area. Outside the metro area, the population is almost precisely stagnant, with a very slight population loss, and that is unlikely to change. Upstate NY overall is very stagnant and stable, and likely to remain that way, without much change. On the other hand, the NYC metro area is much more unpredictable, with the health of Wall Street having much to do with what happens to its growth rate. Given the shape of NY, and its respective regional growth rates, one can predict what happens to one CD, mine, with a pretty high degree on certainty. In that sense, the situation I think is rather unique out there on the Fruited Plain.

I wouldn't leave out the possibility that Columbia goes with Albany after 2020. The Albany metro has to be chopped, and assuming that NY doesn't necessarily hew to the UCC pack rule the chop could give the Capital district a major role in two CDs.

[muon2]

True, but that would be politically toxic. I strongly doubt that will happen. Swallowing Rensselear (sp) is enough. Reaching down to Columbia, which has nothing to do with Albany, and everything to do with the rural Hudson Valley, is something else entirely. And I don't know how such a scheme would affect the chop count, and under the new law, chops matter - all of them, and if the legislature goes rogue, the highest NY court should bounce it.

Some rural counties will invariably be appended to UCCs. Can Columbia make a case that is different from any other county adjacent to the Albany UCC?

As I read the redistricting language, I don't see that chops have that high of precedence. Compactness gets its own line as a principle (who defines it?), but chops are buried with political considerations and CoI.

You must be logged in to read this quote.

[muon2]

It is also interesting to note that one could construe principle 2 to be INEQUALITY, principle 4 to be EROSITY and principle 5 to include my other 3 scores: COMPETITIVENESS, SKEW, and CHOP (with UCC factors). Perhaps my idea to adjust CHOP by adding SKEW and maybe COMPETITIVENESS isn't so far fetched. It would be consistent with treating them as a single related principle.