Just in case anyone is thinking about statehood, I've decided I should make some hypothetical maps for Washington, D.C. and Puerto Rico. (Someday, in September, when we have data.)
NPR reports that the US Census Bureau might not release redistricting data until September 30, long after the usual dates in February through April, or the previous delayed estimate of July. BDistricting's fully automated approach should be just fine though. In a crunch we can rent a bunch of cloud servers and have good results for any state as little as a day after the data is available.
The 2020 Census data for redistricting is immanent starting around 2021-01-19, but we don't have official reapportionment yet! The last report from the Census bureau was "early 2021, as close to the statutory deadline as possible". If they don't have proper apportionment out by the 19th, I guess I'll start my redistricting engine based on estimates (or this estimate, or others, or some consensus I put together).
One of the innovations in these court challenges was the "ensemble" method of producing thousands of plausible maps and doing statistics across likely outcomes of those maps. Kagan found that this demonstrated "the State’s map was an out-out-out-outlier" in the case of North Carolina's map which was worse than all 3000 other proposed maps it was compared to. Roberts mocked this claiming, "Judges not only have to pick the winner—they have to beat the point spread." Roberts at other points wanted to avoid such 'hypotheticals', but Kagan found that "... the same technologies and data that today facilitate extreme partisan gerrymanders also enable courts to discover them, ..." Kagan isn't afraid of technology, it's not some weird magic black box to her. It's a tool, and it's not some weird hypothetical, it's a tool actually used to create real and effective gerrymanders, and the same tools can analyze and pick apart those gerrymanders.
With great handwringing, Roberts repeatedly claimed that there was no solid rule on which to base a decision or decree a change. Kagan responds with a long list of citations of Supreme Court and other Federal Court decisions that found rules and made decrees.
Roberts implicitly suggested that Congress and State Legislatures could make laws creating such rules. He cited a list of bills relating to redistricting rules to which Kagan responded, "what all these bills have in common is that they are not laws. The politicians who benefit from partisan gerrymandering are unlikely to change partisan gerrymandering. And because those politicians maintain themselves in office through partisan gerrymandering, the chances for legislative reform are slight." (Emphasis Kagan's)
Kagan doesn't think we should wait for ballot initiatives either: "Fewer than half the States offer voters an opportunity to put initiatives to direct vote; in all the rest (including North Carolina and Maryland), voters are dependent on legislators to make electoral changes (which for all the reasons already given, they are unlikely to do)."
And on independent commissions, "Some Members of the majority, of course, once thought such initiatives unconstitutional. See Arizona State Legislature, 576 U. S., at ___ (ROBERTS, C. J., dissenting)"
It makes districts. They're pretty simple. Given the alternative of complex gerrymandered monsters simple sounds like a good feature.
My usual criticisms of "process" oriented districting applies. Nothing about this process will necessarily produce "good" districts. Its simplicity ignores all existing county and city boundaries which would make it hard to manage. (Something my own software misses too, which I 'm working on for 2020.)
In the dichotomy of solutions being "fair process" vs "good outcome" this is another "fair process" entry. I like its output subjectively better than the "shortest splitline" algorithm, and the county-cohesion feature will make it more practical, but I still think there's nothing inherent in its structure that will ensure a "good outcome". I still prefer to start at the end, define what kind of outcome we want, and achieve that by whatever methods happen to work. If we have the right metric, then optimizing to it will be okay. (On the other hand, the last century of industrial best practices point out that picking the right metric is problematic and good process needs to be blended with metrics for actual best outcomes.)
tl;dr: they did some weird things I think will have messy unintended consequences
In 2021 they'll appoint a "non-partisan state demographer" who will do the first pass at map drawing. But the law has a pretty strict set of rules to design to.
In order of priority:
- Equal Population
- No racial bias (or could be read to require VRA style gerrymandering-for-good)
- "Partisan fairness and, secondarily, competitiveness". Fairness being: "the difference between the two parties’ total wasted votes, divided by the total votes cast for the two parties, is as close to zero as practicable." Furthermore, competitiveness: "simulate elections in which the hypothetical statewide vote shifts by one percent, two percent, three percent, four percent, and five percent in favor of each party. The vote in each individual district shall be assumed to shift by the same amount as the statewide vote. The non-partisan state demographer shall ensure that, in each of these simulated elections, the difference between the two parties’ total wasted votes, divided by the total votes cast for the two parties, is as close to zero as practicable"
- Respect existing political boundaries