Friday, June 23, 2017

The problem in the first half is the solution in the second half! You can't have it both ways.

From "Inkblot" is the wrong meme for the gerrymandering problem in the case the Supreme Court is looking at. by Ann Althouse. Althouse is a retired constitutional law professor who is accustomed to reading texts in a critical and detailed manner. Most of us vacuum up words like a baleen whale does krill. We don't distinguish the individual words but get a general sense.

Althouse looks at every word no matter what she is reading - a law, an article, a song, literature.

That detailed critical reading sometimes pays off. In this article she shines light on the incoherence of a NYT article, Gerrymandering Case Echoes in Inkblot-Like Districts Across the U.S. by Michael Cooper.

Gerrymandering is a favorite bugbear for reformers on both sides of the aisle, a position I long shared. And still do to an extent. What has changed for me is that I still think gerrymandering to be a vile effort at manipulating power but I am now more inclined to suspect that it is not strategically as consequential as I once believed.

There are three articles of faith in gerrymandering conversations. 1) Gerrymandering is real. 2) Gerrymandering is effective at changing the voting outcomes. 3) Gerrymandering is an evil exclusively perpetrated by the other party.

We know gerrymandering is real though its extensiveness is disputed.

We know that both parties gerrymander.

What has changed in the past twenty years are new questions arising as to whether it makes much of a difference. The gerrymandering forced by the Civil Rights Act has been effective at ensuring greater racial representation in Congress than might otherwise have been the case. However, since the African American vote is overwhelmingly a Democratic Party vote, the Civil Rights Act gerrymandering is simply constitutionally sanctioned political gerrymandering.

While all partisans believes gerrymandering works, there hasn't until recent years, been much of a capacity to test that assumption. However, with the increasing power of computers and access to voting data, researchers in recent years have been able to run simulations that allow us to answer the question about effectiveness.

Obviously any outcome makes a difference to an individual candidate and the voters for that position but that doesn't determine whether it makes a difference in the aggregate.

What researchers have done is used actual party affiliation and voting patterns at a polling station level and run those numbers against randomly generated districts and compared the results. What they have found is that when you compare actual party counts from gerrymandered districts against thousands of randomly generated districts, in aggregate the differences in outcome are minute: 1-5% in most of the research I have seen.

Again, I am not discounting that gerrymandering throws districts towards one party or another and that in particular circumstances a 1% difference can swing things one direction or another. However, the great bulk of the electorate live in largely homogenous areas and therefore only a few voting districts are routinely competitive between the parties and in only a few of those competitive districts does it appear that gerrymandering influences the outcome at all.

We are investing a lot of effort fighting over gerrymandering and yet the evidence seems to suggest that, while real, it simply doesn't influence the outcomes in a predictable or material way.

Which is all backdrop to the actual NYT article Althouse is dissecting.

The NYT is a solidly Democratic Party organization and so in their worldview, gerrymandering is done excessively by Republicans to the detriment of Democrats and they want it stopped. In this instance, Cooper has accidentally gotten himself tangled up. Most criticisms of gerrymandering hinge on the appearance of a district shape - is it an incomprehensible mutant (gerrymandered) or is it compact, following terrain and geography (not-gerrymandered). This is simplistic heuristic as gerrymandering is the manipulation of voter headcount to an predictable outcome. The fact that in order to get the headcount numbers to work the geographical map may look peculiar is incidental to the actual targeted event (headcount manipulation.)

Cooper mistakes the visual evidence of gerrymandering with the actuality of gerrymandering. He starts with an instance in Pennsylvania where there are all sorts of Republican gerrymandered districts with very odd shapes and extended forms.
A Rorschach-test inkblot of a district that has been likened to “Goofy Kicking Donald Duck,” this district meanders through five counties and is so narrow in parts that it is only the width of a restaurant in King of Prussia and of an endoscopy center in Coatesville
For Cooper, gerrymandering in Pennsylvania is bad because it is being done by Republicans and it leads to weirdly shaped districts.

In contrast, in Wisconsin, the gerrymandering is bad because it is done by Republicans and it leads to compactly shaped voting districts.
Indeed, one of the defenses made by Wisconsin officials is that their districts are compact.

[snip]

“They don’t look bizarre,” William Whitford, one of the Democratic plaintiffs suing over the Wisconsin map, said Monday on a conference call with reporters. “But if you really know the Wisconsin political geography — and that’s a learning curve! — they are bizarre.”
As Althouse points out:
Did the NYT not notice that the article is insane? The problem in the first half is the solution in the second half! You can't have it both ways. Which is kind of why the Supreme Court hasn't figured out what to do with these cases (other than to allow the litigation to proceed, which is some sort of deterrent to the most aggressively partisan gerrymandering).
You could easily gloss over the self-contradicting argument Cooper makes. Althouse is much more attendant to the particulars of the argument which is presumably what makes her a good constitutional lawyer.


No comments:

Post a Comment