This time next year we may be very thankful for Republican gerrymandering

[Recursion]

There are two conflicting narratives out there about the gerrymandering Republican state legislators have been engaged in since the last census. One is that the GOP has drawn districts that increase its lead in the House; the other is that gerrymandered districts are making incumbent Republicans' re-elections safe. However, from a math standpoint, only one of those can be true. Fortunately for us, it's the first one.

If you're a state legislator drawing up district lines, you have a finite amount of your party's voters in the state, and you can do one of two things with them. You can spread them into as many districts as possible while still keeping them the majority in those districts, or you can cram them into a district that leans so heavily to your party that a challenge would be futile. Some of both happens in different districts and different states, but the main thrust of GOP state legislatures this decade has been to expand the Congressional majority. And it worked: despite losing the national popular House vote by 1.4 million, the GOP retained control of the chamber.

While effective, this tactic is also very risky. Consider a hypothetical state with 5 districts and a population of 2 million people (just to keep the math easy; actual House districts are somewhat larger). Imagine you're a Republican flak tasked with drawing up the new district lines. You know that you have 2 million people (let's pretend they're all adult registered voters), and you know that 800,000 of them are Republicans (again, just for math's sake). You have to divide up these voters into 5 groups of 400,000. What do you do?

The "safe seats" answer is to draw the lines so that your 800,000 Republicans make up the entire population of 2 districts. A Democratic win in those two districts is now inconceivable. However, you've also left the other three districts entirely in Democratic hands.

The "expand your caucus" answer is to put 200,001 Republicans in four of the five districts (assume you could find the extra 4 people somewhere): you now have a majority in 4 out of 5 House districts in your state. But it's a very slim one on a per-district basis -- two Republicans in each district staying home on election day could cost you all of your seats.

Obviously I'm simplifying here, but the math still applies in real life: shoring up safe seats and expanding your caucus are fundamentally at odds. Every Republican voter that's in a "safe" district is one that wasn't put in a riskier district you're trying to take. And obviously nobody would implement either strategy as extremely as my examples, but the tension is there, and it may well work in our favor in the mid-terms. The "pick-up" districts that got drawn by Republican legislators are inherently unsafe districts for them, because they made so many of them and had to dilute their votes into as small majorities of those districts as possible. The GOP might have been better served by trading some of their House majority for larger majorities in the districts themselves.

[Laelth]

1. Interesting. Thanks for posting. k&r n/t

-Laelth

[randome]

2. Very perceptive.

[hr][font color="blue"][center]Birds are territorial creatures.
The lyrics to the songbird's melodious trill go something like this:
"Stay out of my territory or I'll PECK YOUR GODDAMNED EYES OUT!"[/center][/font][hr]

[muriel_volestrangler]

3. You've simplified too much; this is useless with looking at what they actually did

Yes, they're not so dumb as to put all 800k Repubs into 2 districts. And neither are they so ambitious as to think they could predict the voting to one vote, and organise it so they get 4 seats. You've picked the 2 unlikely scenarios as your illustration. What they have done in reality, because no every Republican is as stupid as Louie Gohmert, is more like put 250,000 Republicans in 3 districts, with 150,000 Democrats; and 25,000 Republicans in 2 seats with 375,000 Democrats. This gives them 3 seats in which they get 60% of the vote, which is pretty safe, and the Democrats get just 2.

Here's the actual results - and look at Pennsylvania as a gerrymandered state: https://en.wikipedia.org/wiki/United_States_House_of_Representatives_elections,_2012#Pennsylvania

D 85.1%; R 15.0%
D 89.4%; R 9.4%
R 54.7%; D 41.1%
R 59.4%; D 34.4%
etc.
In PA12 the R lead is only 3% - that's a winnable seat. The next smallest R lead is 13.2%.

If they really piss off a large part of the electorate, that '60-40' advantage might be overcome.

[Recursion]

4. The math was just to illustrate that the more districts you target, the weaker each one is

Because you have a finite supply of voters. Expanding your district count makes each district less safe. And if this turns into a wave year like 2006, then you can blow your own foot off.

[muriel_volestrangler]

7. 2006, however, was a change against a Repub president exposed as incompetent

To get a Dem majority in the House, you need to win all the seats that the Repubs currently have a majority of 6.6% or less in (from a spreadsheet that gave all the votes for the House). You've got to do that, however, on top of a vote in which Obama provided some 'coat-tails' (as compared to 2004, in which Bush was probably still providing a small advantage to Repub candidates).

2004 House: R 49.4% of the vote; D 46.8%
2006: R 44.3%; D 52.3%

2012: R 46.9% D 48.3%

So the D lead in 2006 was 8%; the 2012 lead of 1.4% needs to get back to that 8% just to get a majority of 1 seat in the House. The chances of anything being "blowing your foot off" - ie a worse result for the Repubs than a purely proportional allocation of seats - are, roughly, zero.

[hobbit709]

5. Throw in all the voter suppression efforts and you created 5 districts they control.

It's not gerrymandering alone. It's all the voter ID and suppression efforts to make sure the D leaning voters can't/won't vote

[Recursion]

6. Voter suppression is a HUGE part of it, probably more important than districting

For that matter, blue states had much higher turnout in 2012 than red states, in no small part because of suppression efforts, but that's a big part of why a 1.4 million vote lead on our part didn't turn into control of the House, too: everybody in MA voted, but the much fewer people in Kansas who voted got to elect representatives, too.

(Watch this conceptual space; I'm doing some analysis on state-by-state turnout in 2012 today...)

[Scuba]

8. Not even close. Your assumption that each district has the same (or similar) numbers ...

... is very flawed.

[Recursion]

9. Census.gov is down, so I'll concede while reserving a right to respond later...

From the data I remember, very few states have vast population differences between their own districts.

[former9thward]

14. You are correct.

Districts within states have populations almost equal. Average district is 711,000. In the small population states there are differences between them. Montana has the biggest (994,000) and Rhode Island the smallest (328,000). http://en.wikipedia.org/wiki/List_of_United_States_congressional_districts

However you are leaving out the Voting Rights Act. Because the VRA requires super majority minority districts map makers are required to bunch Democrats together to make super majority districts (70%) and leaves Republicans to be spread around into 55% districts.

[JVS]

15. I think a lot of people here forget the VRA when gerrymandering is the topic.

An unintended consequence of the VRA is that outside of the majority-minority districts, the entire spectrum is shifted to the right.

[Jim Lane]

11. Same, no. Similar, yes.

I've seen redistricting in which this tool was employed, i.e., districts with larger populations where the line-drawers didn't like how those people voted. There is, however, a limit to how much of that a line-drawer could get away with. Exact numerical equality isn't required, but too great a disparity in district size will support a successful court challenge.

The principal means of gerrymandering is still the strategy identified in the OP -- packing the disfavored party's voters into a few districts where they have an overwhelming majority, leaving your party with slimmer but still comfortable majorities in a larger number of districts.

In the current situation, this can also be the basis for threatening some of the Republicans who, thanks to such gerrymandering, won in 2012 but won by slim margins. They have to worry that a significant shift toward the Democrats will put them out of office. Those are the Republicans who might be intimidated into providing the 17 additional signatures and votes we need to end the foolishness.

[Scuba]

12. Thanks for this clarification. The vast universe of my ignorance is now a tiny bit smaller.

[dsc]

16. within a state they have to be very, very close when originally drawn

now population movement after drawing will make some larger than others, but when drawn the districts are as close to the same size as humanly possible. There is a Supreme Court precedent which mandates that.

[rgbecker]

10. Great insight, Recursion.

How it plays out is still to be seen. Also in play, besides voter suppression, is the effect of the primaried moderate republican. You don't have to go far to read stories of old republicans aghast at their party's turn to the right. Could Republicans actually lose seats as this rightward shift is realized in gerrymandered districts? In 2012, at least two Senate seats were thrown to Democrats because of overly jealous rightwing Republican candidates.

[bemildred]

13. I concur, there will be such a tradeoff.

And it is evident that some considerable effort has been required for some time now to maintain the illusion, and I think it's quite possible 2014 will be a "watershed" election, a new governing coalition really asserting itself for the first time.

[kentuck]

17. I think you may be right.

It's not as much a given as some might think.