Sunday, March 27, 2005

1st and 3rd Districts: Safe Seats

I hate incomplete sets, so I created maps for the 1st and 3rd Districts:

1st District

3rd District

These maps illustrate what a "safe" seat looks like. Neither John Larson (D-CT1) or Rosa DeLauro (D-CT3) have faced a tough election in quite some time (Larson has never had a real challenge to his seat, I'm not sure about DeLauro). This is good for Larson, who was actually beaten by Bill Curry for the 1994 governor's nomination despite being an early favorite (ouch). Indeed, Republicans don't really fund elections in these two districts, and they are completely ignored by the NRCC (National Republican Congressional Campaign).

In both districts, most of the towns that usually vote Republican voted for the incumbent Democrat in 2002 and 2004. This adds further credence to the Incumbent Rule: "All things being equal, voters will trend towards the incumbent." Why else would conservative Granby (which voted for George Bush in 2004 and is represented by Republicans in the General Assembly) go for Larson, one of the most liberal members of Congress, by nearly 17% in 2004?

What's interesting about the map of the 1st District is its unusual shape. I have commented about the odd gerrymandering of districts before, but District 1 is an egregious example. While the combination of Johnson and Maloney's districts into District 5 in 2001 created some uncertainty as to who might win, District 1 (which is largely defined by District 5's borders) was created to reinforce the strength of Democratic candidates there. Why, for example, are the Republican towns of the Farmington Valley (Granby excluded) not included in District 1? Why are small towns like Hartland and Colebrook not in District 5? Why isn't Bristol in District 5? Grographically, this would make a lot more sense.

The 3rd District isn't quite so bad; at least it seems geographically consistent in most areas (the major exception being Middletown). Still, it's obvious that both districts were created with their incumbents in mind.

This can be problematic if said incumbent decides to venture out of the safe confines of his/her district. Take the sad case of Barbara Kennelly, who ran for Governor in 1998. Kennelly had never really faced opposition in the 1st District, and must have assumed that her style of not actually campaigning or paying attention to her opponent would work wonders outside Hartford County. She was, of course, crushed by politically canny Gov. John Rowland (don't blame me, I voted for Groark).

So perhaps the lesson here is that gerrymandered districts make lazy politicians. And who wants that?

If anyone is interested in the Excel files I created to make these maps, here they are:
1st District
2nd District
3rd District
4th District
5th District
Election results for 2004 and 2002 are available on the website of the Secretary of the State of Connecticut:


stomv said...

Loving the maps.

So, I was playing with the spreadsheets, and I "drafted" the following trades:

From 1 to 5:
New Hartford

From 5 to 1:
New Britain

2004 voter differential net change:
8590 voters to District 1.

2004 votes differential:
District 1 gains 30845 GOP votes
District 5 gains 30845 Dem votes

This, of course, is for the 2004 election, and can be used to predict but not guarantee future outcomes. Also, it uses votes and not voting-eligible citizens for the numbers, which are a good but not perfect estimate.

Nevertheless, perhaps Genghis can draw up these two stomvmandered districts. I think you'll find that they look more sane.

You'll also note that Larson won CT-1 by 125,000 votes, so the 31,000 differential isn't likely to hurt him. Johnson won CT-5 by under 61,000 in 2004 but only 23,000 in 2002. A 31,000 vote swing would make CT-5 competitive, but still seemingly likley to remain GOP.

This change would
(a) un-gerrymander CTs 1 and 5 so that they appear much more reasonable in terms of convexity.
(b) keep CT-1 easily Dem, but move CT-5 from GOP to in-play.

It is unclear from this very brief suggestion
(a) if this effects voting-rights Ammendments
(b) if this is kosher in terms of population. It's certainly close, but probably isn't perfect. Chopping up a town along the border could get it done.

Now, I know some folks aren't interested in re-drawing the lines to an advantage. I suggest that this isn't doing that -- rather, it's re-drawing the lines to remove the GOP advantage and level the playing field, taking voting interests into consideration over competitive races. If genghis does the map, it'll be pretty clear that this districting makes much more sense geographically.

Personally, I'd love to carve off bits of CT-1 and send 'em to CT-2 as well. Mostly, I'd love to figure out how to squeeze out Shays in CT-4 in a reasonable way, but other than shifting some of CT-3 to CT-4, some CT-4 to CT-5, CT-5 to CT-1, and CT-1 to CT-3, it doesn't look likely.

It boggles my mind that CT currently has 3 GOP Congressmen. It's clearly a function of the system being gamed, most notably IMO CT-4 and CT-2... since CT-5 is the most compact of the 5 congressional districts. There's simply no reason that CT shouldn't be 3:2 Dem, or even 4:1 were the lines drawn "just-so". As far as I'm concerned, the Dems should try to ram through the changes as soon as possible... why wait until 2010?

Genghis Conn said...

I still hate the idea of redistricting "out of season" (large ethical issues aside, we'd get pasted in the media), but since we're all speculating, why not?

Interesting work, stomv. I checked the population numbers, and they're pretty close. Your plan trades 158,721 to the 5th district from the 1st, and 160,415 to the 1st from the 5th. Districts in CT currently contain around 680,000 people (this is a somewhat high number compared to the national average--and a topic for another time).

The problem is that Nancy Johnson lives in New Britain. She'd be a serious threat to Larson, who couldn't campaign his way out of a paper bag. Let me propose an alternative:

N.B. and Plainville stay in CT-5. Meriden and Cheshire migrate to CT-1. The population swap would be is 86,787 to the 1st from your 5th, 88,866 to the 5th from your 1st.

Total population swap: 158,366 to the 1st from the 5th, 160,800 to the 5th from the 1st. Yikes, math.

As for the map... what am I, a map-making fool? Am I here to make maps, day in and out?

Obviously, yes. It's a good thing these aren't hard to make.

Areas shaded with both green and blue are areas you and I disagree on. Your idea makes more sense geographically, mine takes Johnson's residency into account.

stomv said...

Well, now we're talking. Assuming that NJohnson is still in office in 20xx (next redistricting), than I'd be more than happy to accept your plan. While mine looks "theoretically" cleaner, yours takes into account current residences, and the reality is that redistricting a congresscritter out of his or her own district sure smells like gerrymandering.

So, I'd certainly accept your proposal over the current status quo, even if it isn't as geographically "clean."

I don't think the Dems would get pasted in the media for a mid-decade redistrcting. The GOoPers have done it a few times with little fanfare. The Tejas debacle only got news because it involved 50 Democrats "on the lam".

It looks like the Dems have put all of their eggs in baskets 1 and 3. A redistricting plan that spread them around just a bit might help make CT-2, CT-4, and CT-5 more competitive. Any changes to the districts that are done in a "geographically-improving" way would be hard to fight it would seem.

Genghis Conn said...

Well, CT-2 is competitive, that's why both parties pay so much attention to it (which translates into a lot of obnoxious phone messages left on my machine), and CT-4 is becoming more so. I couldn't say how CT-5 would do without Johnson in the race, but the presidential vote shows that there are a lot of potential Democratic voters out there. Could be close.

I would actually like to see districts redrawn to emphasize parity between the parties. CT-2 is like this, and it always makes for an interesting race. Parity ensures that issues of importance to the district are given plenty of attention. Until recently, Democrats in CT-4 had little chance of their congressman hearing their concerns during an election cycle, because CT-4 was so reliably GOP. The same goes for CT-1 and CT-3 for Republicans.

As for getting pasted in the media... it seems like the kind of thing Bill O'Reilly, Rush and friends would pick up on very quickly.