Introduction
While few people would argue against the proposition that the Supreme Court Justices are some of the most intelligent legal luminaries in the United States, these luminaries often rule in opposing directions. The number of the Court’s split decisions often outnumber the number of unanimous decisions. During the 2021 Term, the number of 6-3 splits alone outnumbered the number of unanimous decisions.
It should also be noted at the outset that the justices’ decisions in all cases before them are foremost decisions on whether to affirm, reverse, or vacate lower court decisions. This post looks at the individual justices’ practices and votes in relation to each of the federal circuit courts of appeal. To do this we must define some measures that allow for this comparison. Before getting to that though, we can see qualitatively, that the justices do not always seem eye to eye.
This dissensus in the justices’ views is sometimes evident in oral arguments. Some justices, for example, may pitch easy questions to attorneys allowing the advocates the ability to expound on their positions in a case. Take for instance the recent oral arguments in Loper Bright v. Raimondo. The differences in Justices Thomas’ and Sotomayor’s questions to attorney Paul Clement who represented Loper Bright illustrate this point.
Early in the argument Justice Thomas asked Clement a fairly straightforward clarification question:
JUSTICE THOMAS: Mr. Clement, you heard the government’s, the General – General’s arguments with respect to the use of mandamus as a basis for sort of deference. Could you comment on that? Because my understanding of mandamus is that a duty has to be clear before it actually lies, but I’d like your comment on that.
Just moments later, Justice Sotomayor asked for Clement’s position much more pointedly and in the process, insinuated her preconception of that line of argument:
JUSTICE SOTOMAYOR: But the point is that I don’t – it’s great rhetoric, Mr. Clement, but we do delegate, we have recognized delegations to agencies from the beginning of the founding of interpretation. And so I – I – I — I’m at a loss to understand where the argument comes from.
Differences in justices’ views are clear elsewhere as well – in particular in the justices’ votes and opinions. In New York State Rifle & Pistol Assn., Inc. v. Bruen, for example, these differences were at center stage. Justice Alito, with his concurrence, spoke directly to the merits of the dissent’s position when he said:
“The dissent makes no effort to explain the relevance of most of the incidents and statistics cited in its introductory section…Instead, it points to studies (summarized later in its opinion) regarding the effects of ‘shall issue’ licensing regimes on rates of homicide and other violent crimes. I note only that the dissent’s presentation of such studies is one-sided.”
The dissenting opinion written by Justice Breyer retorted:
“JUSTICE ALITO asks why I have begun my opinion by reviewing some of the dangers and challenges posed by gun violence and what relevance that has to today’s case. Ante, at 2–4 (concurring opinion). All of the above considerations illustrate that the question of firearm regulation presents a complex problem—one that should be solved by legislatures rather than courts.”
Along with disagreements in terms of legal reasoning and application, the justices often disagree on outcomes. When this is scaled up beyond the individual case level we can pinpoint percentage differences in agreement through the justices’ voting alignments. Last term, for example, Justices Sotomayor and Thomas only agreed with one another 53% of the time while Justices Roberts and Kavanaugh agreed with each other 95% of the time.
Lastly, the justices predominately disagree with the lower courts in cases that the Court takes up. This is evident from the 2/3 or so of the Court’s merits docket where the Court vacates or reverses the lower court decisions. In a select few of these cases, the justices go to an even greater length by explaining the flaws in the lower court decision. One clear example is from the Court’s per curiam decision in 2014’s Lopez v. Smith. Here the Court wrote,
“We have emphasized, time and again, that the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), 110 Stat. 1214, prohibits the federal courts of appeals from relying on their own precedent to conclude that a particular constitutional principle is ‘clearly established…’ Because the Ninth Circuit failed to comply with this rule, we reverse its decision.”
Measurement
How do we generate expectations or probabilities for how the justices will differentially decide in cases based on the lower courts that heard such cases below? The first thing to do is decide on the method. As Bonica and Sen point out, this has historically come down to alignments between the judges’ and justices’ parties of appointing presidents:
“The seminal research of Segal and Spaeth (2002) showed that the party of a judge’s appointing president is a powerful predictor of Supreme Court decisionmaking across a variety of subject matters. Perhaps the simplest way to “estimate” judicial ideology for all federal judges—Supreme Court Justices, federal appeals judges, and federal district judges—is to assign them either the partisan affiliation or the ideology of the US president who appointed them. The simplest operationalization is to compare Republican-appointed judges to Democratic-appointed judges.”
Epstein, Landes, and Posner underscore the importance of this measure in their book The Behavior of Federal Judges,
“The most common ex ante measure of judicial ideology (the most common measure of judicial ideology, period) is the party of the appointing President. It has several advantages over other measures: it is unambiguous, usable for all Article III judges (because all are appointed by the President), highly correlated with other measures of judicial ideology (including ex post measures), and easy to understand.”
Bonica and Sen though also convey weaknesses in this approach:
“Of course, the party of the appointing president is, at best, an inexact proxy of judicial ideology. Judges appointed by presidents of the same party can differ significantly in terms of their jurisprudence and policy preferences; for example, judges appointed by Donald Trump are, according to most reports and related measures, far more conservative than those appointed by Gerald Ford, a more moderate Republican.”
While there is no gold standard for measuring ideology of Supreme Court Justices, and there is even less regarding appeals court judges. Some including Epstein and her collaborators developed a useful approach called the Judicial Common Space (JCS) which bridges Supreme Court Justices’ positions with those of appeals court judges as they describe:
“The goal of our measurement strategy is to place Supreme Court justices and Court of Appeals judges into a policy space that we call the JCS.4 Any measurement strategy that meets this goal should have a number of properties. The measures should be reliable and valid, they should not be issue or time dependent…”
This is the primary measure of interest in the next section. One drawback to this approach though is that while it standardizes positions across time, it does not account for case selection. With this in mind, this method is useful in understanding the justices’ positions relative to one another when they voted in the same cases, but not necessarily between terms. While Supreme Court Justices’ JCS Scores are based on their voting alignments, appeals court judges’ scores are based on the perceived ideology of their appointing presidents and home state senators. While this is also useful in distinguishing between the different circuits, it is a somewhat crude proxy measure for their views in actual cases.
Application
First, some baselines worth noting. This analysis uses three terms of data – from the 2020 through the 2022 term. To gain a better understanding of the justices’ positions relative to the circuits we can first look at the rates of petitioner wins (as coded in the Supreme Court Database) or when the lower court decision is not affirmed. [Note that this is a different measure than pure reversal rates used in other studies]
With this we see that the First and Ninth Circuits’ decisions, for instance, are more frequently overturned than those from the Federal or Eleventh Circuits. The Court, however, does not review decisions in equal amounts from all circuits and oftentimes the number of cases the Supreme Court hears from the various circuits correlates with the circuit sizes. Also, it is important to note from the graph above that the Supreme Court generally takes cases that the justices wish to reverse so even with the justices’ vote in cases from more liberal or conservative appeals courts based on the judge composition of the circuit, a commonality is that the justices still take most of these cases with an eye towards reversal.
Here are the case counts by circuit for the three-year period focused on in this post:
These numbers, and in particular the disparity between cases from the Ninth Circuit and those from any of the other circuits should not be surprising given the size of the Ninth Circuit and the greater number of decisions coming from that circuit than from any other circuits.
Finally, we have some prior conceptions of how the judges may vote relative to one another and relative to the circuits based on JCS Scores:
Since the JCS is not updated on an annual basis, the numbers are from the most recently released data which includes the justices’ scores for 2020 Term and the numbers for the circuits from the year 2022. Also, since the appeals court judges scores are based on appointing president and senators and not on actual votes, the difference in years between the two measures should not harm this analysis. As we are only looking at the differences between justices and between circuits, the bridged aspect of the scores between the Supreme Court and the circuits holds only minimal value here. The ordering of justices and the appeals courts from the most to least liberal is the primary area of focus. Lastly, the numbers for the circuits are based on their median scores or the score for the middle judge in each circuit based on ideology. This is why the range of values for the circuit court medians is much tighter and closer to zero than the justices’ scores.
We can generate expectations with these data about the difference in votes of more liberal and conservative justices based on whether the judges in a circuit are coded as more liberal or conservative. Justice Sotomayor, for example should theoretically side more often with the First and D.C. Circuits than Justice Thomas who we’d expect would vote to uphold more decisions from the Fifth and 11th Circuits than Justice Sotomayor. This, of course, assumes that the Court is not only taking highly liberal lower court decisions from all of the circuits for review.
The data for each justice shows their rates of affirmances for cases from each circuit as well as how frequently each justice was in the majority in cases from the various circuits. Let’s begin with the Court’s liberal justices. While percentages are generally used, Justice Jackson’s analyses include case counts because she is the only justice who was not on the Court for all three terms (this is also evident in the number of cases listed as “Not Participate” in her graphs).
We can see differences even within the liberal justices. Justice Sotomayor, for example, voted to affirm 75% of the decisions from the Fourth Circuit while Justice Kagan voted to uphold only 12.5% of cases from the Fourth. Justice Jackson was much closer to Justice Kagan as she voted to overturn all Fourth Circuit cases before her. All three liberal justices voted to predominately uphold decisions from the 11th Circuit and all three had high rates in the majority for those cases. For Fourth Circuit cases though, Justice Sotomayor was less frequently in the majority than Justices Kagan and Jackson. All three justices had low rates of voting to affirm the (recent) frequently analyzed Fifth Circuit decisions although Kagan voted to uphold these decisions more often than the other two justices.
As we see from these justices, there is not always a strong correlation between the perceived ideology of the circuit and the justices’ ideological positions. The 11th Circuit is the most conservative circuit according to the JCS, but the three more liberal justices all had high rates of affirming that circuit’s decisions. It is also worth noting though that the justices’ percentages are more likely to fluctuate significantly if the Court took few cases from a circuit during this period.
Now for the six conservative justices from most to least conservative according to the JCS:
Here we see that the conservative justices also all had high rates of affirming 11th Circuit decisions. In fact, all justices aside from Sotomayor voted to uphold 11th Circuit decisions most of the time leaving almost all of them almost frequently in the majority in those cases.
Several of the conservative justices had high overturn vote rates for the more liberal, D.C. Circuit’s cases, although Gorsuch and Thomas voted to uphold those decisions more frequently than the other conservative justices.
We also know from these data that Justice Kavanaugh’s positions most highly correlate (among the justices) with the Court’s decisions in these cases as he was in the majority nearly 100% of the time if not 100% of the time for decisions reviewing cases from each of the circuits.
While most of the justices had relatively high rates of upholding decisions from the Federal Circuit, Sotomayor was a clear outlier in that respect. According to the JCS, the Federal Circuit is one of the more liberal circuits and this might explain why Justice Sotomayor had a particular affinity for those decisions.
Even with the supposition that the Fifth Circuit has turned highly conservative since Trump appointed the six judges, most of the conservative justices still voted to overturn a high number of Fifth Circuit decisions. The more conservative justices tended to uphold higher percentages of Fifth Circuit decisions. Justice Alito was the only justice to vote to uphold more than 50% of Fifth Circuit decisions (54%) and Justice Thomas had the next highest percentage voting to uphold about 46% of the Fifth Circuit decisions under review. Justice Kagan was not far behind though at approximately 41%.
Most justices voted to overturn the often perceived liberal Ninth Circuit’s decisions much more frequently than they voted to uphold them. Still, the two most conservative justices, Thomas and Alito, were the only conservative justices that voted to uphold Ninth Circuit decisions more than 20% of the time. On the liberal end, Justice Kagan’s vote percentage in Ninth Circuit cases was in the same range as Justices Alito and Thomas, while Justices Sotomayor and Jackson were much closer to 50% rates of voting to uphold decisions from the Ninth Circuit
Concluding Thoughts
This term the Court has granted 61 cases for argument. At the high end, 12 of these cases come from the Ninth Circuit, ten from the Fifth Circuit, and seven from the Second Circuit. All other circuits will have fewer cases reviewed. The Court is just past the halfway point in its argument calendar for the year. With only three cases already decided, the Court voted to overturn the First Circuit’s decision in Acheson Hotels, to overturn the Second Circuit’s decision in Murray v. UBS Securities, and to affirm the Third Circuit’s decision in Dept. of Agriculture Rural Development v. Kirtz. All three were unanimous decisions. As the justices decide more cases this term, we will see how well the statistics in this post forecast the justices’ votes.
In terms of the reliability of the analysis and the predictive power of the statistics, some of the presumptions from the justices’ perceived ideologies held true, while several others did not. Does this show flaws in the JCS or in our understanding of the justices? Not necessarily. To reiterate an earlier point, the justices vote to reverse most lower court decisions they review. They tend to cherry pick these cases from each circuit that fit this paradigm so that even with more conservative circuits, the Court might take predominately liberal cases and reverse at a higher rate than expected.
The three years of data only allow for limited inferences and more accurately provide a recent snapshot of the relationship between the justices and the circuits. Additional years and more data could both help and hinder our understanding of the justices’ decisions. The courts of appeals judges on a circuit fluctuate more often than the Supreme Court Justices if not for any other reason that there are many more of them and so there are many more opportunities for turnover. It is hard to draw conclusions over broad time periods for this reason. Fewer decisions over a smaller number of years though might not provide as robust support for any conclusions. These three years only provide a small snapshot of the Court’s cases. Justice Jackson was only on this Court for one complete term and so her statistics are even less telling than those for the other justices. All of this should be accounted for in drawing inferences from the data.
Find Adam on X/Twitter and LinkedIn. He’s also on Threads @dradamfeldman and on Bluesky Social @dradamfeldman.bksy.social
Empirical SCOTUS 