pageok
pageok
pageok
Ninth Circuit Judges Wary of Split:

The National Law Journal reports from the Ninth Circuit Judicial conference that many of the circuit's judges do not like proposals to split up the circuit.

The Los Angeles Times op-ed that has some judges grousing was written by Vanderbilt University Law School professor Brian Fitzpatrick, who was also a law clerk to 9th U.S. Circuit Court of Appeals Judge Diarmuid O'Scannlain of Portland, Ore., an outspoken supporter of a circuit split.

Fitzpatrick suggested that "it can be shown mathematically that as a court grows larger, it is increasingly likely to issue extreme decisions." . . .

Of the potential for "extremist" decisions in a large circuit, Chief Judge Mary Schroeder said, "You have got to be kidding. We don't appoint the judges, the president does. You don't split up a court because you don't like the decisions it makes." . . .

Schroeder noted in the circuit's annual report issued this week that, at the peak of congressional efforts last year to split the 9th Circuit, 33 of the 47 total active and senior judges signed a statement of opposition to a split.

bittern (mail):
A pet peeve of mine is judge types who underestimate their scientific ignorance. Vanderbilt University Law School professor Brian Fitzpatrick, an outspoken supporter of a circuit split, suggested that "it can be shown mathematically that as a court grows larger, it is increasingly likely to issue extreme decisions." . . .

Does Fitzpatrick know what he's talking about? What is his basis for this statement? What Wikipedia seems to call the law of large numbers would suggest the opposite -- that the more voters/judges, the closer the vote would adhere to the mean.
7.27.2007 10:48am
Cornellian (mail):
I'm guessing Fitzpatrick means the more judges on the court, the more likely you will have at least some judges with outlying views, and the more likely it is that random coincidence will give you a panel of 3 outlier judges, hence "extreme" decisions.

However, to say that this necessarily entails that 47 is too large seems to me to be completely arbitrary.
7.27.2007 11:06am
Cold Warrior:
bittern, Fitzpatrick's argument is different. Because the 9th Circuit pulls 3-judge panels from a pool of 47 active and senior status judges (and sometimes even reaches beyond that, with district court judges sitting by designation), Fitzpatrick points out that the chances of pulling an "extreme" panel (at least 2 judges at the far "right" or "left" of the spectrum) are greater in the 9th Circuit than in other circuits.

My problem is that the effect is very small. I would call it negligible, but Fitzpatrick (looking at it from a different angel -- the huge number of cases decided by the 9th) thinks it matters. And he has a good point.

I think there are better arguments for splitting the 9th, but Fitzpatrick's is clever and worthy of discussion. It is not worthy of dismissal with a wave of the hand in the manner of Judge Schroeder.
7.27.2007 11:09am
bittern (mail):
Thanks, both, for the clarification. You might think so, but I believe statistically that wouldn't increase the likelihood of extreme results, compared to a panel from a smaller pool, unless by some indirect effect like the nominator feeling free to pick more extreme people for a larger panel (mistakenly thinking the outliers would wash out).

I'm assuming the panels are picked randomly. If not, my math is pretty irrelevant.
7.27.2007 11:31am
AnonLawStudent:
Bittern,

Fitzpatrick has put forth his numbers for all the world to see. Do you have numbers, or do you just "believe" in a different result?
7.27.2007 11:56am
bittern (mail):
A very fair admonishment, AnonLawStudent!

His math makes his point because he puts three extreme judges in each half. If you knew you wanted to employ a known set of extreme judges, then you could put exactly one in each circuit and two would never double up. If you don't know who your extremes are in advance, the size of the court doesn't matter. Once you find out who is who, then you can split people you don't want together. If you split his model court randomly, there's a good chance you'd have more than three extremes in one squad.
7.27.2007 12:24pm
NaG (mail):
I have a question. If the Ninth Circuit is fine at the size it is, then why not merge some other Circuits together so that every Circuit is of similar size? What is the advantage to having one Circuit be larger than all the rest by a significant margin?

Assuming that the administrative weight of having a Circuit be larger is not an issue, then what's the compelling reason for having so many Circuits? Heck, why not have only two? Or even just one?

Seems to me that having multiple Circuits helps assist the Supreme Court in identifying important areas of law that warrant resolution -- through "Circuit splits." If this is the case, then does the Ninth's size assist in this? (My guess: no, since it's large enough to be split into another Circuit that can identify its own splits.) Would it be better to have fewer Circuits? (Again, no, for the same reason.) Or are we better off having Circuits that are more similarly-sized? (No, yet again.)
7.27.2007 12:26pm
Fub:
Cold Warrior wrote:
My problem is that the effect is very small.
From TFA:

Fitzpatrick suggested that "it can be shown mathematically that as a court grows larger, it is increasingly likely to issue extreme decisions."

Fitzpatrick said that on a court of 27 active judges -- the current number serving on the 9th Circuit appeals court -- it is fairly likely that with just six judges holding "extreme" views, two would be randomly selected for a three-judge panel. But in a split circuit of just 14 judges -- and just three "extremists" -- it less likely that two extremists would be selected for a three-judge panel.

OK, let's do the numbers.

Hypergeometric calculator is here.

Select 3 judges from:

Current panel:

Population = 27, #extremists = 6. What is probability of making 3 draws w/o replacement yielding at most 1 extremist? Result: .88547

Split panel:

Population = 14, #extremists = 3. Probability in 3 draws w/o replacement where at most 1 draw is extremist? Result: .90659
7.27.2007 12:30pm
Steven Lubet (mail):
Why assume that extremists are wrong, or that extreme decisions should be avoided? After all, those extreme judges (whoever they are) were appointed and confirmed just like the others -- why shouldn't their votes occasionally count?

Moreover, sometimes extreme decisions turn out to be enlightened or pathbreaking, so the optimum number of extreme decisions must be >0.
7.27.2007 12:50pm
bittern (mail):
If you halve a well-shuffled deck of cards and then pick five cards randomly, your chance of getting three kings is the same as if you hadn't halved the deck before picking. But it's different if you get to pick which cards go into each half of the deck. You could reduce the chance of multiple extremes in the 9th if you pay attention to who went into each half. But it's not really the size of the court that causes the reduction in multiples, it's all in the picking.

You'll note that the Congress doesn't pick members of its conference committees randomly.
7.27.2007 12:52pm
Just Dropping By (mail):
Steven Lubet has an excellent point -- "extreme" decisions by circuit courts (in the sense of one circuit reaching a completely different result than multiple other circuits) are sometimes ultimately found to be the "correct" result by the Supreme Court. Ashcroft v. Free Speech Coalition is a good example of this -- four other circuits (1st, 4th, 5th, 11th) had upheld the CPPA as being constitutional and the Ninth was the only circuit to strike it down. In the end, the Supreme Court sided with the 9th's analysis.
7.27.2007 1:30pm
TerrencePhilip:
The clue to the fact that the 9th is out of step with the Supreme Court is the high number of unanimous reversals of 9th Circuit opinions. But it appears that all or almost all the far-left judges are in California. So by splitting the 9th you could well create one circuit with an even higher reversal rate, and another with a very low one.
7.27.2007 2:45pm
BruceM (mail) (www):
Forget about reversal rates, "extreme" opinions, and other such considerations.

Is there any reason NOT to split up the 9th circuit and make half of it the 12th Circuit solely to make its size more in-line with that of other federal circuit courts of appeals? It must be hard to administer a court as large as the 9th. It has so many more judges and covers so many more states than all the other circuits. I'm not talking about quality of judicial output. Ignore that, because that's a political issue, ultimately (i.e. whether or not you agree with the output). I think california and 3 other states from the 9th circuit should be put into a new 12th Circuit (or keep those states in the 9th and put the others in the new 12th Cir.).

It took a lot less to split the 5th Circuit up into the 11th, and I have not heard anyone regret that decision or otherwise argue that it was not efficient.

As with the 11th circuit vis a vis 5th circuit precedent, the new 12th circuit will begin with all 9th circuit precedent, not a clean slate. So nobody should object on political grounds, either.
7.27.2007 3:24pm
Spartacus (www):
One reason extreme decisions are also more likely in the 9th is because their "en banc" rehearings do not include all the judges of the court; it's really just a bigger panel. In another circuit, an extreme panel is more likely to be reversed by an en banc panel of the entire, presumably more balanced court. But in the ninth, you may never get that chance, because your en banc panel may consist of one wing of the court as well.
7.27.2007 3:57pm
Just Dropping By (mail):
It took a lot less to split the 5th Circuit up into the 11th, and I have not heard anyone regret that decision or otherwise argue that it was not efficient.

Splitting the 5th was easier, however, because of the presence of two large states to anchor each of the resulting circuits. The population of all the other states in the 9th Circuit don't add up to California's.
7.27.2007 5:31pm
David M. Nieporent (www):
If you halve a well-shuffled deck of cards and then pick five cards randomly, your chance of getting three kings is the same as if you hadn't halved the deck before picking.
But that's clearly not the case. On average, there are only two kings in each half. If there are, you can't draw three.
7.27.2007 5:38pm
David Stras:
I am a big fan of Brian's, as I said on Empirical Legal Studies, but I disagree with him on this issue. His numbers, first, assume that the "extreme" judges will be split equally between the circuits. But my guess is that Brian would argue that most of the extreme judges are in California. If that is so, it is far more likely that four or five extreme judges would end up in the "California-plus" circuit, which makes it equally or more likely that an "extreme" decision will be rendered by a panel of two or more extreme judges. Check out the following post on SCOTUSblog for more on the issue.
7.27.2007 6:01pm
David Stras:
I might add that Frank Cross was the first to point out the problem with Brian's assumptions in the comments to a prior Volokh Conspiracy post.
7.27.2007 6:05pm
bittern (mail):

But that's clearly not the case.

David Nieporent, if I lay the shuffled cards out, in order, in four columns of thirteen, and you hope to get at least three kings among your five picks, I will enjoy watching you strategize about which rows and which columns to pick the cards from. (Cuz I'm like that.) For the non-superstitious, it doesn't matter which ones you pick.
7.27.2007 6:23pm
bittern (mail):

I might add that Frank Cross was the first to point out the problem with Brian's assumptions in the comments to a prior Volokh Conspiracy post.

And rather more concisely than I.

How could we statistically test whether two identical topics would grind into the same rut on the Volokh?
7.27.2007 6:32pm
David M. Nieporent (www):
Bittern, I don't know why you think that's responsive to what I wrote. I didn't say anything about which of the two was more likely than the other to come up with 3 kings. I said that either of the two is less likely than both of the two combined.
7.27.2007 9:35pm
OrinKerr:
Steve Lubet:
Why assume that extremists are wrong, or that extreme decisions should be avoided? After all, those extreme judges (whoever they are) were appointed and confirmed just like the others -- why shouldn't their votes occasionally count?
I think the assumption is that the "extremist" judges aren't making much of an effort to follow the law, and that judges should be following the law.
7.28.2007 12:34pm