Supreme Court Roulette
How to make the Supreme Court a better place
People have lots of ideas for improving the Supreme Court. Cameras in the courtroom. Term limits. Congressionally-imposed ethics rules. Expanding the Court. Forcing Justices to ride circuit. None of these ideas will ever happen, but we debate them anyway.
We’ve been thinking too narrowly.
I’d like to propose a new idea that has numerous virtues and, as far as I can tell, no flaws whatsoever. I call it Supreme Court Roulette. It works like this:
The Justices do what they currently do—vote at Conference and write their majority, concurring, and dissenting opinions.
When the opinion is ready to come out, one Justice is selected randomly. That Justice’s view becomes the opinion of the Court.
So if there’s an 8-1 decision, there’s an 8/9 chance that the majority’s view will become law and a 1/9 chance that the dissent’s view will become law.
By the end of this post, you will be convinced. I promise.
(I don’t know how Dall-E came up with that funky robe, but I like it.)
Great minds do not think alike
Why do we have nine Justices? Why not just one?
It’s because the authority to be the final arbiter of all disputes of law, for life, is too much power to vest in one person. Justices stay on the Supreme Court for decades. To have a single person lord over the judicial system for such a long period would transform that person into Louis XIV.
When there are multiple Justices, disagreement is inevitable. We need a convention to determine whose opinion prevails. That convention is: majority rules.
But why should the majority rule?
One possible justification is that if more Justices believe a legal conclusion is right, it’s more likely to actually be right. If we assume that legal problems have objectively correct legal answers, then, ceteris paribus, we’re more likely to land on the objectively correct answer if we take the answer that more Justices believe. If you give nine equally distinguished mathematicians an extremely difficult math problem, and seven say the answer is “4” and two say the answer is “3,” then it seems more likely the correct answer is “4” than “3.”
I can hear you all snickering.
In the real world, decisions that get more votes aren’t inherently better-reasoned. The three most prominent Supreme Court decisions of last term—303 Creative v. Elenis, Students for Fair Admissions v. Harvard, and Biden v. Nebraska—all featured identical 6-3 lineups, with the Republican appointees in the majority and the Democratic appointees in dissent. One cannot infer that the majority decisions are better-reasoned merely because more Justices voted for them; one can simply infer that there were philosophical differences between the Republican appointees and the Democratic appointees, and the Republican appointees prevailed because there were more of them.
Of course, many decisions have more unusual lineups. But even in those cases, it’s impossible to say that decisions that got more votes are “better.” All nine Justices are extremely intelligent people. If a single Supreme Court Justice thinks a particular legal conclusion is correct, that alone is a sufficient basis to pronounce that conclusion, at a minimum, reasonable. If that Justice is outvoted, it reflects a difference in philosophy, not a logical error that the dissenting Justice was unable to discern.
Supreme Court Roulette would make opinions different, but it wouldn’t make them worse.
I guess it’s legitimate the other 99.98% of the time?
Another justification for majority rule is that Supreme Court Roulette seems arbitrary and illegitimate. Why should the law of the land be determined randomly?
If you think Supreme Court Roulette is arbitrary or illegitimate, I have news for you. The overwhelming majority of cases in the federal system are decided in exactly this way.
In general, when you file a case in federal district court, your case is randomly assigned to a judge in that district. For example, if you’d like to file a lawsuit in the Northern District of New York, your suit will randomly be assigned to one of the five active judges and four still-sitting senior judges. One randomly-chosen judge from among those 9 will decide your case. Supreme Court Roulette works the exact same way, except that the Justices get the benefit of conferring and sharing ideas before the judge is randomly selected.
Even if you then appeal, there is a lot of randomness. Appeals from the N.D.N.Y. go to the Second Circuit. The Second Circuit has 18 Democratic appointees (7 active and 11 senior) and 11 Republican appointees (6 active and 5 senior). Setting aside variations in how often the judges sit, there’s a 68.4% chance that the panel will have a majority of Democratic appointees. Under Supreme Court Roulette, there’s a 66.7% chance that the randomly-chosen Justice will be a Republican appointee. It’s basically the same system, except there are more Democratic appointees on the Second Circuit than the Supreme Court.
In 2022, there were about 380,000 cases filed in federal district courts across the country. Last term, the Supreme Court issued 58 merits cases. So more than 99.98% of cases don’t go up to the Supreme Court and are decided based on Lower Court Roulette. In a way, it’s weird that this doesn’t happen at the Supreme Court.
Even at the Supreme Court, there is considerable randomness as to which Presidents appointed a majority of Justices at any given time. Without relitigating the Garland, Gorsuch and Barrett nominations, I think we can all agree there are many worlds in the multiverse in which there isn’t a 6-3 majority of Republican appointees.
We cannot expunge randomness from judicial outcomes. We can only manage it.
Everyone should win sometimes
Supreme Court Roulette wouldn’t make outcomes worse and it wouldn’t make them any less arbitrary. Now that we’ve dispensed with all possible arguments against Supreme Court Roulette, let’s turn to the arguments in favor of the proposal.
Let’s adopt an extremely crude model in which all Supreme Court cases are either (a) easy cases, resulting in 9-0 decisions, or (b) hard cases turning on philosophical differences, wherein all Republican appointees vote one way and all Democratic appointees vote the other way.
Supreme Court Roulette wouldn’t change the outcome of the easy cases. As to hard cases, however, under the Supreme Court’s current configuration, Republicans would win two-thirds of the hard cases under Supreme Court Roulette, as opposed to 100% of the hard cases in the status quo.
So which system is better, majority rule or Supreme Court Roulette?
Right now, I’m sure judicial conservatives prefer majority rule and judicial progressives would prefer to win sometimes than never. But behind the veil of ignorance, where you don’t know whether judges who share your philosophy are the majority, which system is better?
Supreme Court Roulette is better. The aggregate output of the Court’s decisions would be more reflective of the country than majority rule. In a divided nation, we want Democrats to win sometimes and Republican to win sometimes, rather than Republicans or Democrats to always win (or Republicans to always win for a few decades, and then Democrats to always win for a few decades, depending on which party can eke out a majority at any particular time).
In addition to the aggregate output of the Court’s decisionmaking being more moderate, individual decisions would also be more moderate. Why? Because Justices would have to treat all outputs of Supreme Court Roulette as binding precedent. If there are precedents going in both directions that must be followed, decisions will necessarily have to accommodate both sides’ views. And moderate decisions are good—if the Justices are going to be bitterly divided, decisions may as well reflect the views of the median voter.
This would lower the national temperature. Individual Supreme Court nominations wouldn’t be as dramatic because the new Justice could only possibly affect the outcome one-ninth of the time and because decisions—both individually and in the aggregate—would be narrower. Isn’t it nice for everyone to be friends?
Also, people can’t get that mad about Supreme Court Roulette. It’s hard to get mad at randomness. It’s like getting mad at the weather. It’s just one of those things. After the outcome of a big Supreme Court case is randomly determined, Republicans and Democrats can go together to the bar.
No more Kennedy Briefs
Of course, this extremely crude model doesn’t reflect reality. Not all Republican appointees invariably vote for the conservative position, and not all Democratic appointees invariably vote for the progressive position. Indeed, in recent years, a pattern has emerged wherein a “swing” Justice decides virtually all big cases. For many years, this was Justice O’Connor and to some extent, Justice Kennedy. When Justice O’Connor retired, Justice Kennedy became the undisputed solo swinger. There are no swing Justices like those anymore, although some Justices today are closer to the center than others.
We have nine Justices because it’s bad for one Justice to decide every important case. It’s therefore bad when eight Justices are evenly divided 4-4 in every important case, causing one Justice to … decide every important case. Who made Justice Kennedy the King?
Also, when Justice Kennedy decided all big cases, lawyers would write briefs exclusively for Justice Kennedy, which could get frustrating for those who did not share his particular jurisprudential views.
Under Supreme Court Roulette, this never happens. Supreme Court Roulette distributes power evenly. Every Justice has the same amount of power as any other.
No more strategic alliances
Supreme Court Roulette would change the Justices’ incentives for the better.
Let’s start with merits-stage incentives. In the status quo, when a Justice has 5 votes, she has no incentive to try to get 9.
Not so under Supreme Court Roulette. The Justices are going to hate picking the outcome randomly. They’re going to do everything they can to reach unanimity.
That’s good! It’s nice when the Justices can achieve consensus. And unanimous decisions have more legitimacy—isn’t it a good thing that Brown v. Board of Education was 9-0? True, unanimous decisions are sometimes garbled given that the Justices have to paper over their disagreements, but a little bit of ambiguity seems better than Justices firing bombs at each other.
Lots of cases yield dissents now, but if the Justices tried harder, perhaps they could get to unanimity more frequently. After all, jurors are required to be unanimous. Occasionally there are hung juries, but usually jurors are able to achieve consensus if they deliberate long enough. Why not demand the same of the Justices?
In some cases, to be sure, consensus will be impossible. The Justices have irreconcilable views on some issues, and dissent is inevitable.
Supreme Court Roulette is better in that situation, too. Under majority rule, the Justices have an incentive to attempt to cobble together a majority, yielding some strange compromise solution allowing a Justice to count to five, with outraged dissents from the other four. When that happens, you’ve got the worst of both worlds. You’ve got the garbled, unprincipled opinion designed to get a majority, and you have Justices firing bombs.
Under Supreme Court Roulette, there’s not much of an incentive to get to 5, because getting 5 votes versus 4 votes only slightly increases the chances of a particular outcome becoming law. And so all the Justices will simply write opinions expressing what they actually think. The opinions will be more forthright, because no strange compromises are needed. And the final decision will at least reflect what someone believes—as opposed to what no one believes.
The opinions will be clearer, too. Cobbled-together 5-4 majorities, fending off outraged barbs from dissenters, are often hard to follow. Under Supreme Court Roulette, you’ll at least get one person’s candid, uncompromised view.
And what about when, under majority rule, the Justices can’t cobble together a 5-4 decision?
That’s the worst of all! You have no idea what the opinion means. Can anyone offer a clear account of what the Dormant Commerce Clause requires following the Court’s 2-2-1-4 decision in National Pork Producers v. Ross? Not a problem under Supreme Court Roulette—you never have to worry about the Marks rule again. Just read the opinion of the Justice whose name is picked out of a hat.
Make the Supreme Court boring again
Supreme Court Roulette would also improve incentives at the certiorari stage. The Supreme Court is prone to taking controversial culture-war cases that seek primarily to prove a point rather than resolve a bona fide legal dispute. 303 Creative, with the designer of roofing websites fending off the fake gay couple, is a classic example of this phenomenon, but there are many others.
Why did the Republican appointees vote to take the case? Because they knew they had the votes to rule for the plaintiff. They would never have taken the case if they knew there was a one-third chance that Colorado would win.
Under Supreme Court Roulette, the Supreme Court won’t take controversial cases unless it really has to. The Supreme Court will grant certiorari in (a) boring cases that are likely to be 9-0, and (b) controversial cases that have sufficient practical significance that a Supreme Court decision is essential. Contentious but fake cases will disappear from the docket. This seems fine to me.
Overruling and overruling
What happens when Justices dislike a decision so much they want to overrule it? Are they just going to keep granting certiorari, over and over, until their preferred position wins?
Unlikely. In the status quo, the Court rarely grants certiorari to decide whether to overrule a case. Under Supreme Court Roulette, it will do so even more rarely, given that it won’t be able to predict the outcome in advance. Supreme Court Roulette will therefore increase fidelity to stare decisis.
Steelmanning the arguments against my proposal, perhaps there are some issues that are so controversial that Justices will keep trying to overrule old decisions, over and over again. It takes four votes to grant certiorari. Suppose there are four Justices who strongly want to overrule Dobbs and four Justices who strongly want to keep it. If the Justices in the former category have sufficient antipathy towards Dobbs, they might keep granting certiorari until they finally win Supreme Court Roulette. If there were 4 or more votes on the Court to keep Dobbs, those Justices might keep granting certiorari until Dobbs was un-overruled. This is a recipe for chaos and social conflict.
I’m not sure I have a good answer to this point, except to point out that there’s chaos and social conflict in the post-Dobbs status quo. It’s not clear Supreme Court Roulette would make the situation worse.
You see! I promised you’d be convinced.
Kudos to Mr. Unikowsky for his excellent and detailed proposal to correct the many issues with the Supreme Court by using a random process to select a single Justice who would author the controlling opinion in a case. But (why is there always a ‘but’ after a compliment) there are some good arguments against the proposal which ultimately render his proposal inappropriate.
1. There is his statement “All nine Justices are extremely intelligent people.” Oh God, that this were true. In Clarence Thomas we have a Justice who is by qualifications, temperament and intelligence uniquely unqualified for the Court. He sits on the Court only because of his prior political affiliation and the color of his skin. He has never authored a significant opinion, in part because it seems likely that every CJ has been told “Never give Clarence the opportunity to author a major opinion, he doesn’t have the slightest idea of what he is talking about.”
If Clarence had his way states would be able to imprison gay men and women for the sole reason that they were gay men and women. The Establishment Clause would only apply only to the federal government, state would be free to establish a state religion. Connecticut would be free to ban the sale of condoms. A person facing capital punishment whose innocence was proven would still be put to death if his lawyer failed to adequately represent him or meet arbitrary deadlines for various filings. Well you get the picture.
The post states “Supreme Court Roulette would make opinions different, but it wouldn’t make them worse.” Justice Thomas is living proof that is not the case.
2. As a retired quantitative method professor, I can tell you that a random procedure does not always produce what looks like a random outcome. Here is a random number sequence taken from a table of random number. 1688722228. That’s right, out of a run of 10 numbers, 8 are even and the number 2 appears consecutively for four times.
Want another example. Including the 1992 election there have been 8 presidential elections. The Democrats have won a plurality or majority in seven of these. Democrats have been President for 20 of the 32 years since 1992, Republicans 12. Dems 62.5%, Republicans 37.5% of that time span. The Dems have appointed 33.3% of the Justices.
Randomness is a bitch.
3. One argument that is used in the post is that at the District level and the Circuit level judicial decisions are made by one individual judge chosen randomly. This is not entirely correct. If a District judge makes an error, there is the Circuit judge who can make a correction. A Supreme Court decision is final unless Congress acts to overturn it, an extremely unlikely, probably impossible action in today’s partisan divided House and Senate.
If the Circuit judge makes an incorrect decision there are two remedies. The first is a hearing en banc which can overturn the decision, the second is an appeal to the Supreme Court. While both of these remedies are rare given the number of cases that go through the federal court system, they do exist and they do happen. This would not be the case with a single Justice rendering the opinion.
Finally, we have judge shopping and single judge Districts/Divisions. If there is a specific outcome one wishes and that one person is a conservative, well, I got a place in Texas just for you. No randomness filing there. Need the appropriate jurisdiction for your case, no problem.
4. There is a question of when should the selection of the single Justice be made. Make the selection after oral arguments and you have wasted the time of 8 of the 9 Justices.
Make it before oral arguments and 8 out of 9 Justices do not have to prepare or participate in oral arguments. The job of Supreme Court Justice is already a part time job. This change would make it a part time part time job.
5. One reason that the disparity in the Court vis-à-vis Republican vs. Democrats is that people do not play fairly. Two of the Justices appointed by President Trump were legal appointments that violated the unwritten rules and decorum of Supreme Court nominations. Where power is concerned we cannot trust anyone to not violate the rules when those rules are not absolute law.
So under the single opinion author, randomly selected system there may be nothing to prevent Justice A who feels very strongly about the issue from cutting a deal with Justice B who won the opinion lottery but does not care to write about the issue. Justice A could say I will trade you my next two opinions if you let me author this one. If this was prohibited by law the Justices would simply resort to clandestine procedures. There is no honor among thieves or ideologues of any ideology.
6. The Court, like the nation is bitterly divided along extremes. But a majority of Americans like the middle. We want equal opportunity, but without quota’s or excessive preferences. We want parental rights in education, but not have parents dictate school curriculums. We want abortion to be safe, legal and rare and we want the welfare of the mother to be protected, but not have late term abortion on demand. A Supreme Court which is constructed of extremists on both sides would further divide the nation if a single Justice controlled an opinion.
So when the post states that “We cannot expunge randomness from judicial outcomes. We can only manage it” that is not correct. In fact, we cannot manage randomness. That is the nature of randomness. When we try to manage randomness the outcome of that effort is almost always to make the situation worse. To paraphrase Churchill, the process of deciding Supreme Court issues by majority rule is terrible and awful, except compared to all the other methods.
This is fascinating, the most intriguing court reform idea I have seen floated in recent memory. I hope you will follow up and address some critiques. I want to believe in this idea, but have some skepticism.
1. This seems to work really well at the level of the judgment, but maybe not at the level of reasoning and how far the decision goes? In Dobbs, there were 6 votes to uphold the law at issue and 3 to strike it down, so under your system the dissenters would have had a 33% chance of prevailing instead of 0%. But if I am Clarence Thomas, why don't I, instead of joining the Alito majority, simply concur in the judgment while writng an opinion that - like his Dobbs concurrence - sweeps more broadly and strikes down Obergefell, Lawrence, etc, figuring that such an opinion does not reduce the odds of reversing Roe but does give an 11% chance of getting everything I want even though 89% of the court does not agree?
2. It does seem like, on political cases, conservatives (because they have the majority now) might keep taking bites at the apple until they finally prevail.
3. Would this approach be strictly for merits opinions or also apply to emetgency motions?