While everybody has the supreme court on their minds, I will avoid talking about this week's issues while they are so raw. Instead, I want to look at the issue of politically biased courts and what we might do about it. (I have had the distinction, unusual for a foreigner, to be involved in 2 cases before the U.S. supreme court, once as an appellant, and once as chair of the legal group representing an appellant.)

In short, can we design voting rules for a court that would reduce its ability to be biased? There are several popular proposals, including term limits, or viewing 5-4 rulings as a tie, which might help things.

Courts are going to be appointed by politicians, and as long as the judges can make rulings for political reasons, that leaves us with biased courts. But ideally, judges on a court are not biased, and their rulings are independent of any preconceptions, both political and legal.

Forbid correlation

In statistics, we have ways to test if things are independent events. We can measure their correlation. We could, if we wished, make rules forbidding too much correlation between judges. If a group of judges are regularly voting together, they are not being independent jurists, and this can be curtailed.

Let me explain with an example rule for illustration that is probably much too simple:

The justices of the court can never repeat a vote pattern within 20 years

This rule would mean that once the judges had voted judge A,B,C,D on one side and judge E,F,G,H,I on the other side, they could not rule in that pattern again for some period of time. If there is a group of 5 on one side and a group of 4 on the other, they could issue one 5-4 ruling, but they would have to pick it well, because they could not do that again. If they met after a case and found they wanted to repeat a ruling, somebody would have to change, possibly by switching to abstention, or even changing sides. In the event of a deadlock on changing (which is probable mostly on 5-4 splits) another formula, ranging from something based on past history to a simple random choice, would break the deadlock.

A 6-3 majority would get its biased way much more often -- 48 times in fact -- so the rule could be tuned to limit that in a different way.

These rules would eliminate not just political leanings. They restrict all correlations between the judges. Even ones on apolitical judicial theory (rare as that is.)

Since judges change, the rule has to account for that. At it simplest, each replacement would wipe out the old history and start a new one. Alternately, the replacement could be considered a continuation of the record of the prior judge. That works in the more common replacement with a judge of similar leanings. In a complete flip, it would mostly reset the history. The history could also "age" so that patterns from the past get lower or even zero weight, and patterns across a judge replacement age-out even faster.

Will this cause bad rulings?

Some correlations will be just coincidence, or what we might view as less harmful biases. As such, this system might force rulings to change even when there is not a bias problem. I hope that could be tuned to minimize such things, but I feel that the elimination of the large biases is probably worth some cost. As I have noted, the very simple rule would not be the real rule, and something with more statistical sophistication would be used.

How to write the rules

A simple rule has the advantage that the public can understand it. But the rule does not have to be simple. Even if a complex statistical algorithm that is best done on a computer is used, we all have computers, and we all have an intuitive feel when there is too much correlation. Statisticians have all sorts of great models on how to measure correlation and quantify it. If the computer tells the court that a ruling pattern is too close to past rulings, they can act from that without calculating it themselves. The computer can also offer variations of the desired ruling which are valid for their deliberations on how to change.

Could this actually happen?

Well, to make it a hard and fast rule would require a constitutional amendment, probably. That's unlikely, but the one hope is that since the rule would be completely objective and mathematical, it will not be partisan. The party that currently holds "power" on the court might fight a rule that eliminates the idea of power, but if they are smart, they might realize it's better to take the power from everybody since down the road the power will shift.

There have been some suggestions of non-amendment ways to change the court. The constitution says very little about the court, other than judges are on it as long as they have "good behaviour." It doesn't say how many there are, or how it makes decisions. It's not impossible congress could declare that decisions have to be 6-3. It's more likely though, that people would view this (or any formula above) as a violation of traditional separation of powers.

It's been proposed that term limits could be imposed by the senate or President demanding of any nominee that they pledge to resign after 20 years. Don't pledge, and don't get on the court. The problem, of course, is that when the Senate and President are aligned politically, they want to put their judge on the bench for as long as they can. In an adversarial situation (which tends to produce far less partisan judges) they could demand the term pledge, and perhaps make it much harder for the next judge to not be asked.

I am not kidding, by the way, when I say that the term problem has to be solved because I expect medicine, some time this century, to solve the problem of aging. Quite possibly within the lifetimes of Gorsuch or Kavenaugh. If so, these men have received quasi-forever appointments to the court, which is not what is desired. In this case, a constitutional amendment would become much more clearly needed. At that time, other variations could also be considered for the amendment. The ability for the court to be partisan offers short-term gain, but long term divisiveness for the country.