Gerrymandering: A Rant

First off, foundationally, districts are reconfigured within a US state once a decade, after the most recent census. Doing it at any other time, for any other reason than "we now have better data on how many people live where," is inherently illegitimate.

Which, of course, is why Texas Republicans are doing it - they've never seen a naked power grab they didn't like. If it spreads, and lots of states do it as well, that will be yet another breakdown of norms, another signpost that the US is speed-running through the Collapsing-Democracy Playbook. So that is bad, obviously.

But there's a lot of talk essentially assuming gerrymandering is magic - that just drawing the boundaries some secret way ensures the people drawing them will win. And that's not the case. 

The other incorrect assumption is that gerrymandering creates "safer" districts: ones more ideologically committed to the governing party. That it tends to reinforce the echo-chamber effect, to lock in ideologically extreme candidates. Again, that's not true.

Gerrymandering - as we're discussing it right now, meaning an effort to maximize the number of districts in a state that are favorable to your party - tends to make districts less safe, on purpose, to make more districts that a particular party has a solid chance of winning.

The opposite strategy is to make districts where your party can't possibly lose, where the voters break for your party overwhelmingly. Democrats have tended, over the past few decades, to create that kind of district, at least in some cases - usually as "majority-minority." (And, yes, that will and has tended to dilute their overall statewide voting power - yet another example of how Dems will always bring a flower to a knife fight. Also, yes, there are other voting systems besides first-past-the-post, and it's really nice that you support them, but you do realize Texas is never, ever going to spring for approval voting, right?)

Especially in larger states, both strategies can be in play - a governing party can "pack" a few districts with their opponents, maybe also "pack" one or two on their side to protect specific incumbents, and then "crack" the rest of the map into districts that are mildly in their favor.

But remember that, no matter what you do in a gerrymander, there's still the same number of voters overall - what gerrymandering does is turn a state with (to make up numbers) twenty districts from this evenly balanced list:

District	Partisan Lean
1	R + 10
2	R + 10
3	R + 10
4	R + 10
5	R + 10
6	R + 10
7	R + 10
8	R + 10
9	R + 10
10	R + 10
11	D + 10
12	D + 10
13	D + 10
14	D + 10
15	D + 10
16	D + 10
17	D + 10
18	D + 10
19	D + 10
20	D + 10

to this list, where one party is dominant in more districts:

District	Partisan Lean
1	R + 10
2	R + 5
3	R + 5
4	R + 5
5	R + 5
6	R + 5
7	R + 5
8	R + 5
9	R + 5
10	R + 5
11	R + 5
12	R + 5
13	R + 5
14	D + 10
15	D + 10
16	D + 10
17	D + 10
18	D + 10
19	D + 10
20	D + 10

But note that the totals - my example uses partisan lean as the metric; in the real world it's actually the number of voters, and gets very data-wonkish very quickly with race/gender/propensity overlays - are the same in both cases. Gerrymandering does not invent voters, it moves them to be more efficient for the party in charge.

Gerrymandering is a bet - that you'll take a somewhat smaller chance of winning each of a larger number of districts, and that you can do so while keeping the partisan lean higher than the potential swing in any given election.

Because, in the state above, if "partisan lean" has a 10-point error bar, every district could potentially have been a toss-up in Scenario 1, and Scenario 2 would give the R team a chance in any given election to shoot themselves really badly in the foot.

But if the error bar is more like 3 or 4 points, a +5 district is pretty much optimal for your side. You'll win those nearly every time, and any higher lean in your direction is "wasted."

(Note: I am not a data scientist and this is not my area of expertise. I make no claims about what the typical error bar is, or the partisan lean of existing or proposed districts in any state. I do expect the experts know these things, and make their plans based on decent data and reasonable assumptions.)

The more aggressive you go in gerrymandering, the more likely you make it that any given election - possibly the very next one - will shift enough in the direction of your opponents that you will lose power.

This nuance tends to get forgotten in discussions of gerrymandering: that the more you push it, the higher the chances it will go wrong. Yes, we do have really powerful computers, robust data models, and all kinds of insights, but it all still relies on what a large number of people will do on a specific day in the future, so wave elections - and we've had a succession of wave elections for more than a decade now - can be enhanced by an aggressive gerrymander.

Now, Texas Republicans have other tools besides gerrymandering, of course - have we mentioned intimidation and voter suppression lately? - so I'm not trying to say it's sunshine and roses. They do intend to make it harder for groups they think will vote against them to get to the polls, and shifting districts is good cover to do that. (If the gerrymander goes through, expect polling places to tend to move toward red areas and mysteriously disappear or lose hours in blue areas.) And if those ICE raids sweep up citizens, or make citizens less willing to put their necks out, that's good for their side as well. It is, and I don't want to minimize this, a comprehensive "flood the zone with shit" plan, with multiple levels of bad actions.

All I wanted to point out is that gerrymandering, itself, is just one piece of that plan, and that basic arithmetic applies to gerrymandering. It doesn't break the laws of physics; it doesn't invent voters. And doing it really aggressively because one fairly stupid and narcissistic guy insists does not necessarily mean they will win all of those districts in the next election.