PPIC generally, and Eric McGhee specifically, largely do good work. So it’s not surprising the statistical work in their recently released, “Assessing California’s Redistricting Commission” is mostly solid. It confirms what close observers know, claims that the commission was “tricked” into passing a Democratic gerrymander are simply not substantiated by those pesky things known as facts.
That said, as someone who closely followed the commission process, when it comes to policy recommendations the report falls short. First the report argues the Commission should consider political data with the explicit purpose of understanding the political impacts of their proposed lines. That purpose would be to look for bias in plans and/or a lack of competitiveness.
The author argues the law’s prohibition against using such data to favor or discriminate against any incumbent, candidate or party does not preclude this. This legal interpretation is questionable. Say the commission drew a district based on the completely partisan-neutral criteria in the law (i.e. compactness, avoiding splitting jurisdictions). Then it tested that district and for whatever reasons it was less competitive than some other configuration. Changing the district to the “fairer” configuration would by definition be making the district better for some candidates and parties and worse for others. You could argue that is a good policy outcome, but it’s still explicitly using political data to change political outcomes.
For better or worse, this is not how Proposition 11 which established the Commission was written. Competitiveness IS NOT a criteria. The goal was not to make California campaigns more interesting for academics, entertaining for reporters, and profitable for campaign consultants. The law says take the politics out of the decision making and then let the chips fall where they may.
You can argue that is naïve. Indeed, states working on redistricting reform proposal right now are struggling with whether to emphasize non-partisan PROCESSES or non-partisan OUTCOMES. Should they emphasize copying California's example or protecting against Wisconsin and Pennsylvania's? Indeed this is often a key difference between favoring an “independent” or a “bipartisan” commission. But California voters made their choice and to suggest simply circumventing it is…. problematic.
More importantly, the author’s argument ignores what introducing such explicitly partisan considerations would do to the commission’s actual function. One of the challenges under Proposition 11 is its requirement that separate majorities of Democrats, Republicans, and Independents on the Commission all vote for the final plans. One of the most critical parts to California’s success was having Commissioners slowly develop a comradery. Over many months they increasingly saw themselves as Commissioners first, and representatives of their partisan pools less and less.
When outside forces tried to argue about the partisan impacts of this line or that line Commissioners were largely able to brush those aside because they knew why THEY drew those lines. They knew they did not do it with a partisan purpose, and therefore they knew they did not have to worry about the partisan outcomes.
The report's proposal would be terribly disruptive to that gravity pulling the Commission together. Imagine every time a map is generated Commissioners getting a report with the results of an opaque statistical calculation of the potential partisan tilt of a map. It would be only natural for commissioners to start looking at whether specific maps were better or worse for their partisan pool. Human natural is human nature, and by making partisanship a part of the Commission’s deliberation, the proposal would push Commissioners to stay in their partisan bunkers. This would make reaching consensus more difficult and increase the chances the commission could fail… all to address a problem the author’s own work suggests is small at best.
This is equally true when looking at the report’s other big proposal: introducing automated line drawing into the redistricting process. Now don’t get me wrong, these computer algorithms can be fun. And they can even serve a purpose, helping to illuminate what is and is not possible with the lines. They can also show how emphasizing one criterion over another changes how a plan comes together.
But from a public policy standpoint, one thing they are lousy at is dealing with communities of interest. If one reads the ballot arguments and explicit language of Proposition 11 this was one of its primary goals, to keep communities from being unfairly divided. There is no central database of communities of interests. Indeed many communities of interest overlap. There is no easy way to tell a computer this community is this much more important to keep whole than that community. It requires debate and discussion, trial and error.
Even if such software did exist, its use ignored another success of the 2011 process. At the time more than a few redistricting “experts” squawked about how the Commission’s demographers should just “tell them what to do.” Even now, I occasionally hear proponents of Proposition 11 talk about how the staff should have reined the Commissioners in more and not let them waste time on things they “obviously” were not going to do.
The problem with those arguments, and with McGhee’s proposal, is they again ignore the human factor. They ignore Proposition 11 fundamentally requires fourteen commissioners (who by law must come from very different backgrounds) to come together to reach a consensus.
One of the brilliances of the Commission’s professional line-drawers in 2011 is they actually let the Commissioners draw the lines. They did everything they could to avoid staff capture (or in the case of the PPIC proposal, algorithm capture). They let the Commissioners explore all their ideas, even the bad ones. This meant later in the process, when tough final decisions had to be made, Commissioners could feel confident in them. They knew that they had explored all the options. They knew when there were less than ideal lines for some communities they were the result of tradeoffs in other parts of the map. They knew they were not the result of staff trying to manipulate them or ignoring creative solutions in favor of those “obvious” ones.
This again insulated the Commissioners against accusations of partisan bias. It may have proved the adage about making sausage. It may not have been fun for outside observers. But to fix the “problem” would only create much bigger issues.
This doesn’t mean tools like the author suggests can’t or even shouldn’t be a part of the 2021 redistricting process. But they should be the purview of outside groups, not the Commission itself. Outside groups can (and I assume will) analyze partisan impacts of lines. When they find evidence of a clear (if unintentional) partisan tilt they will let the Commission know. My strong suspicion is the Commission will respond (see the changes made to draft Board of Equalization boundaries in 2011).
I’m also sure outside groups will submit plans drawn exclusively by computers using this algorithm and that algorithm. And I’m sure the Commission will look at those just like all other publicly submitted maps to understand what is possible and what is not. The Commission will compare them against the public testimony about what should be and what should not.
But to introduce proposals like those in the PPIC report would endanger the fragile balance established by Proposition 11 and California’s voters.