Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

Attending oral arguments in Harris v. Arizona Independent Redistricting Commission

December 20th, 2015, 11:39pm by Sam Wang

sketch by Art LienDec. 25th: Updated with five more FantasySCOTUS predictions. The median outcome still favors the Commission, by a narrow 5-4 majority.

In my NYT piece proposing statistical standards to detect partisan gerrymandering, I focused on Harris v. Arizona Independent Redistricting Commission, a current Supreme Court case [SCOTUSblog] [Blog for Arizona]. On December 8th I attended oral arguments for this case [Supreme Court transcript and audio]. Bottom line: a majority of justices seems disposed toward the Commission, an outcome that I argue is consistent with the data. However, it is not clear whether they will cite any statistical standards.

Oral arguments can help predict how the Supreme Court will rule. Here is my reaction based on (1) what I saw and heard, and (2) what successful Court-prognosticators think.

First off, an aside (this will become important): on the  day I attended, Scalia started the morning by reading out a unanimous opinion on Shapiro v. McManus, a technical but important election law case. Shapiro, who did his own legal work, got his desire to contest, before a 3-judge panel, Maryland’s Congressional districting plan.

Next, the Court took up the Harris case. The Arizona Eagletarian has a good summary which accords closely with what I saw. Counsel for Harris’s side were peppered with questions, first from Justice Kennedy. Kennedy asked whether counsel expected the Court to overturn the lower court’s factual finding that the Commission was motivated mainly by a desire to meet the requirements of the Voting Rights Act. Since Kennedy is the likeliest fifth vote for either side, this was not a great sign for Harris. Throughout the hour of argument, Kennedy needled Harris’s supporters a fair bit.

To my disappointment, nobody asked statistical questions. Ginsburg pointed out that Republicans won a slightly higher fraction of the seats (57-60%) than their share of statewide voter registration (54%). She said that if this was a gerrymander, it was a terrible one. Laughter in the courtroom ensued. A few minutes later, Alito countered that proportionality was not to be expected in a single-member district system, and therefore was irrelevant. In my mind I shouted “Yes, this is why you should quote my amicus brief.” (Note: actual shouting is not allowed in Court.) Alito did not follow up with a more sophisticated statistical point, presumably because doing so would essentially support Ginsburg.

I found, at best, ambiguous indications that any justice would side with Harris. Here are some of the more salient questions.

Scalia: Other than some overpopulation in a few districts, what other evidence do you have? He asked this question of several counsels in a skeptical tone. The replies basically restated the same evidence prefaced by the phrase “overwhelming statistical evidence that.” Based on these weak responses, I could imagine Scalia voting with the Commission.

Roberts and Alito: Are you going to make us figure out, by reading tea leaves, which actions by the Commission were necessary for preclearance by the Department of Justice? That is a mysterious process. The implication here is that it is asking too much to get the Court to engage in such activity.

Scalia again: Is partisanship a necessary evil, or is it just evil? He was bothered that any partisanship was involved at all. (Breyer later pointed out that this was an unrealistically high standard by which to judge the Commission’s work.) Scalia expressed regret that this case did not precede last year’s case challenging the very existence of the Commission. If Scalia votes with Harris, he might have to say that partisan motives must be completely absent. That would involve overturning the existing principle that partisanship in districting is kosher.

Roberts: Where is the district in which, ­­or the state in which, partisanship does not play a role in redistricting?

Kagan, Sotomayor, Breyer: Meh, the Commission stayed within existing rules.

Thomas: As the press likes to write, “He asked no questions, as is his custom.” However, he is quite the reader. Big piles of volumes in front of him.

Overall, it felt like the Commission could get as many as 7 votes (liberal wing plus Kennedy, Scalia, and maybe even Roberts). If I were Harris I’d be unhappy.

If I am wrong, it’s because of the partisanship issue. This has been analyzed in The Economist, which notes questions on this topic from Scalia, Kennedy, and Alito. Also see the writeup at Blog for Arizona. Counsel for the Commission, Paul Smith, a veteran of gerrymandering cases, was dismissive of this angle, as was Sarah Harrington from the U.S. Solicitor General’s office. Harrington pointed out that partisanship is kosher, based on Gaffney v. Cummings. Are Smith and Harrington correct?


The wisdom of the (expert) crowd. I’m far from the only close watcher of oral arguments. Among the most accurate Court-watchers are not professional lawyers or reporters, but fanboys and fangirls: the denizens of FantasySCOTUS, a prediction market in which individuals attempt to predict individual opinions and overall outcomes. Some – who are often not lawyers – do very well. What do they think?

Among top 20 prognosticators on the leaderboard from the 2014-15 term (“OT 2014”), four nine have made predictions on Harris since oral arguments: Bill CortealJosh ErismanKeith Wilson, Tim Delaune, Melech, Justin Abbasi, Jay Medina, nbcrcc, and Scott Meiner. None are perfect: in the 2014-15 Court term, they predicted between 65% and 75% of outcomes correctly. At the high end of this group, Justin Abbasi, predicted 55 out of 73 outcomes correctly. Anyway, all four of them seven out of nine predictions give the Commission prevailing. Of those seven predictions to affirm the lower court ruling, four predict a 5-4 split (in each case, liberals plus Kennedy), two predict that Roberts will join too in a 6-3 decision, and one predicts a unanimous 9-0 outcome. In the remaining two predictions, Kennedy votes with the four more conservative justices. In short, everybody thinks Kennedy will vote with the majority, whichever way things go.


Irrespective of how Harris v. Arizona Independent Redistricting Commission turns out, can a standard for partisan gerrymandering ever be adopted? At least two other possibilities are alive at the moment.

First, in Wisconsin a case is currently under consideration in a lower court. Nicholas Stephanopoulos and Simon Jackman have offered an analysis based on the idea of an “efficiency gap,” a measure of wasted votes.

Second is the Shapiro v. McManus case, which I mentioned earlier in this post. That case now goes back to a 3-judge panel, which will examine the question of whether a gerrymander took place. To many observers, the Congressional district map looks pretty bad – but boundaries can meander for other reasons, so this is a tough argument to make. Using my statistical standards, I have demonstrated partisan asymmetry in Maryland (see my law-review article, at page 33, in Table 1). That looks like a good case for applying these ideas.

Tags: Redistricting

7 Comments so far ↓

  • bks

    Do the experts have a prediction on when SCOTUS will give their decision?

    • Sam Wang

      Too meta, I think. One thing is generally known: more controversial decisions tend to be saved for the end of the term (for example, King v. Burwell).

  • Amitabh Lath

    Sam, they did not have a discussion of statistics because it would quickly get beyond the math skills of either the justices or counsels.

    And let’s face it, the math skills of the reporting classes are abysmal.

    • bks

      It’s only now, nearly half a century after it first perplexed me, that I can appreciate why Statistics is a separate department from Mathematics in the typical university.

  • Arizona Eagletarian

    I, too, was surprised nothing was said during OA about the statistical standard issue you brought up, Sam.

    But I was also concerned in AZ Legislature v AIRC last year when nobody brought up the concept of Republicanism as it was understood in the 1700s. But Ginsberg magnificiently spelled out that very issue in the majority opinion, even if not calling it that.

    My point is that I wouldn’t be surprised if your theory is mentioned in the majority opinion in Harris.

    If I were a betting man, I’d predict 7-2 in favor of the Redistricting Commission. But I’ll be satisfied with 5-4, if it comes to that.

Leave a Comment