Thursday, January 11, 2018

The future of partisan gerrymandering

Big things are afoot here, and I cannot confidently say what the results will be.  And... next week, I gotta lecture on this stuff!

Anyway, some definitions are in order.  First, Elbridge Gerry.  Hard-g, in "Gerry," so hard-g in "gerrymandering."  I'm a pretentious and pedantic asshole about everything.  To goo-goos, gerrymandering is gerrymandering, but... no.  Not all gerrymandering is alike.  Partisan gerrymandering is quite different from bipartisan gerrymandering.  Partisan gerrymandering is drawing district lines in order to spread one party's voters out more efficiently than the other party's voters, thereby allowing the advantaged party to win more seats than its proportion of the population.  It follows the pack-and-crack model.

Observe.  Suppose there are 33 voters, with 21 Democrats and 12 Republicans, to be spread over 3 districts.  If we draw two districts with 6 Republicans and 5 Democrats each, and then the remaining district with 11 Democrats, the first two districts are majority-Republican.  The GOP can win two out of three districts, even though they have only a third of the population.  They do this by "packing" the Democrats into that third district, and cracking the remaining Democrats in the first two.  There are 10 Democrats total in the first two districts-- enough to win at least one district, if they were distributed more efficiently-- but if they are cracked, that doesn't happen.  The pack-and-crack plan distributes one party's voters more efficiently than the other party's voters, and the advantaged party gets more seats than it should because of that.

If you want proportionality in representation, that's bad.  You know what kind of redistricting plan promotes proportionality?  Bipartisan gerrymanders!  That's when you draw a bunch of homogeneously Democratic districts and a bunch of homogeneously Republican districts with no competitive districts.  You also get representatives who are closer to their median voters.  That's why I like bipartisan gerrymanders.  Not all gerrymanders are the same.

Partisan gerrymanders are getting challenged in the courts, but finding a constitutional basis to oppose them is... tricky.  Why?  'Cuz the Constitution doesn't really say jack shit about this.  I don't like partisan gerrymanders, but... there's lots of shit I don't like.  That doesn't make such things unconstitutional.  There's a pending case at the US Supreme Court about this, and the courts have been chipping away at partisan gerrymanders because racial gerrymanders are kind of unconstitutional, and... race and party are kind of linked!

And a North Carolina plan was just struck down by a federal court for being a partisan gerrymander.  The courts are not consistent on this, and that means we need a federal ruling on partisan gerrymanders.  Right now, there is no clear federal standard on partisan gerrymandering.  The Court has said that, in principle, there might be such a thing as going too far, but they have never said how far that is, and a federal court just said North Carolina went too far.  So...  how far is too far?  What's the rule?

Look, this is tricky.  The basic problem here is that nobody likes partisan gerrymanders except the specific party benefiting from the specific partisan gerrymander in question.

But the Constitution just doesn't say jack shit about it!  It wasn't until Baker v. Carr 1962 that the Supreme Court said that districts had to have equal population.  The Constitution doesn't say it!  And, since each state gets two Senators regardless of population... yeah, it's unfair if one district has 1,000,000 people while another has 1,000, but you have to be willing to read something indirect into the 14th Amendment, and it took until 1962 for the Court to do that.

So, the 14th Amendment.  Partisan gerrymanders are unfair, yes.  But a 14th Amendment case against them is... weird.  One party is advantaged over another...  Frankly, if you take this argument seriously, you have to come back to the Senate again anyway.  With Baker v. Carr, the Court basically said that the Senate was written into the Constitution as an exception, and they have to keep doing that, but there is a basic mathematical problem with any district-based electoral system.  All we can ever do is approximate proportionality.

If anyone actually takes this shit seriously, they should advocate eliminating district-based elections and moving towards proportional representation.  Some do.  That's... not what's in the Constitution, and therein lies the core of the problem again.  The basic constitutional structure of the legislative system is one in which proportionality is mathematically thrown off by the electoral system used.

Part of that is that the Constitution was a compromise and a kludge.  Part of that is that they had no intention of thinking about things in terms of parties.  Part of that is that they had no clue about the mathematics of partisan aggregation, but that wouldn't have mattered if they didn't intend parties anyway.  Still, the idea of trying to graft some rule against partisan bias onto a mathematical system that necessarily imposes at least a small bias with a Constitution that doesn't mention any of this at all is...


None of this is to say I advocate partisan gerrymandering.  I'm pretty clear on the record here.  I advocate bipartisan gerrymanders, not partisan gerrymanders.  How will the North Carolina decision play into the Supreme Court's upcoming case?  What will SCOTUS do?  Fuck if I know.  All of this is kind of silly, though.  Within a district-based system, the best we can do is a bipartisan gerrymander, and let's be clear on this:  that means making sure we don't have competitive elections.  Fuck those things.  Making a constitutional argument against partisan gerrymanders, though?  That's a weird and difficult thing to do.  Not every bad thing is unconstitutional, but if anyone really, seriously wants to push lack of bias, that leads either to bipartisan gerrymanders... or a total constitutional overhaul, and PR.  That latter thing... ain't happenin'. 


  1. To come back to a point you made a long time ago about the DeLay Texas redistricting back in the Aughts: how distorted are the results from the statewide balance?

    In the case of NC, it's 10-3 (77% R). From a state that has voted 50.1% R in the two presidential elections held since 2010. If we include 2008, it drops to 49.9%, as compared to 47.3%/48.1% for the Ds.

    But, once we've opened up that can of worms, we've established that you're a prostitute, and we're just haggling about price.

    1. I'm not defending the NC plan. From a mathematical standpoint, the GOP basically pushed it as far as you can go without what Grofman and Brunell call a "dummymander." I actually crunched the numbers for how bad the bias can get for a piece in the CWRU law review, and yes, the NC plan goes pretty much that far. I'm not defending it on principle. I just have a hard time seeing the constitutional argument. Yes, I make a distinction between bad and unconstitutional.

    2. Oh, no doubt. Didn't mean to imply that.

      Just noting that, once we open up the question of partisan gerrymandering, we're in slippery slope territory.

      Which isn't to say it isn't a necessary thing. It just seems likely to take us into a realm of really terrible and ugly Supreme Court decisions.