Justices of the Supreme Court rarely give public comments on words or actions of members of the elected branches of the federal government.  This made the recent spat between Chief Justice John Roberts and President Donald Trump, when the two intrinsically debated the role of partisanship in the federal judiciary, all the more surprising and powerful. While phrased in terse sentences (Trump’s were via Twitter), their words left much to unpack. What began as President Trump decrying the decision of an “Obama judge” (a judge appointed by President Obama), Judge Jon Tigar from the Northern District of California who ruled against the administration’s policy of attempting to keep migrants at the border from applying for asylum, turned into a clash of sorts with the Chief Justice directly responding to the President’s remarks for the first time.

Chief Justice Roberts retorted, “We do not have Obama judges or Trump judges, Bush judges or Clinton judges. What we have is an extraordinary group of dedicated judges doing their level best to do equal right to those appearing before them.” Roberts’ statement then closed with, “the independent judiciary is something we should all be thankful for.”

While the Chief Justice’s statement is admirable and his sentiments are predictable, his analysis overstates the lack of interpersonal differences in the federal judiciary. This Chief Justice, more than most going back to the early days of the federal judiciary, is keenly aware of the importance of its legitimacy.  While in the years prior to Chief Justice John Marshall the federal judiciary was described in Federalist Paper #78 as, “beyond comparison the weakest of the three departments of power,” this characterization is much more debatable at the present.  The Court is often described as activist for ruling on issues that could potentially find resolution through the elected branches of government.  Rather than constrain the Court through, through its own gridlock Congress is one of the great sources that fosters the Court’s political role.

The debate in the current instance though is about what drives judges’ decisions. Are they driven by political views? The judiciary was conceived with the notion of independence from political persuasion. Looking back at the Federalist Papers which convey much of the commentary behind the edicts found in the Constitution, the paramount importance of an independent judiciary is clear. In Federalist Paper #78 we have the words, “If, then, the courts of justice are to be considered as the bulwarks of a limited Constitution against legislative encroachments, this consideration will afford a strong argument for the permanent tenure of judicial offices, since nothing will contribute so much as this to that independent spirit in the judges which must be essential to the faithful performance of so arduous a duty.” This notion of insulation through lifetime appointments was designed to prevent the influence of politics on federal judges as Federalist #78 continues, “That inflexible and uniform adherence to the rights of the Constitution, and of individuals, which we perceive to be indispensable in the courts of justice, can certainly not be expected from judges who hold their offices by a temporary commission.”

As Chief Justice Roberts clearly channeled this philosophical milieu in his comments, he neglected a more realistic assessment of the relationship between politics and the judicial branch. The nomination and confirmation of federal judges has become more politicized over time with the most glaring case coming in the most recent nomination and confirmation of Brett Kavanaugh to the Supreme Court. For those who value the independence of the judiciary, the hope is that once confirmed, judges will strive to make decisions above the political fray (hence the Chief Justice’s words “the independent judiciary is something we should all be thankful for.”)

The truth is much more complicated that this intention though as politics does in fact play a role in judicial decision making. This of course comes down to how we define the relationship between politics and judges’ decisions as well as how we define the influence of politics. On that note we might say that preferences that are actually proxied by politics are highly predictive of judges and especially Supreme Court justices’ voting decisions.  Studies of the relationship between political parties and judges’ decisions go back more than a half century.  Most show the same outcome – that many judges’ votes can be predicted by their political party affiliation or from the party of the president who appointed them. If this weren’t the case then we should not be able to determine how judges will vote based on the party of their appointing president. This should be even more highly underscored in complex cases where judges’ decisions are not clearly dictated by black letter law.  We can infer this complexity through the justices’ rigorous case selection process that focuses on those which have led to disparate results between different judicial circuits or regions of the country.

The 2017 Term Case Pipeline

The notion of federal judges’ votes as divorced from politics is more idealistic than realistic.  Case in point – we can look to the Supreme Court’s decisions across the previous term (2017).  This past term was the first since prior to 2005 where we saw conservative victories in all 5 to 4 ideologically split votes and was the term for the same period with the highest percentage of conservative victories in all of the Court’s 5 to 4 decisions (ideological and non).

Taking a more detailed look, the present analysis examines the 2017 term case pipelines to see how partisanship helps us predict how judges’ vote in complex cases. To do this we can break down each case into its various levels moving from district courts, to courts of appeals, to the Supreme Court. As only federal judges are appointed by the president and are given lifetime tenure, cases moving through state court systems do not aid in this venture and so they were removed. The methodology is simple. We take the party of the appointing president of the district court judge and compare this to the dominant party of the appointing presidents of the three-judge appeals court panel.  We then compare the compare the dominant party of the appeals court panel with the dominant party of the Supreme Court’s majority based on the justices’ appointing presidents. If partisanship affects the judges’ decisions then we should see a higher than random accord between judges’ votes at the three levels when the partisanship matches and a disproportionate discord between judges’ votes when the parties are not matches.

Take for example Epic Systems v. Lewis, argued on the first day of oral arguments for the 2017 term.  The flow of this case moved from the District Court for the Western District of Wisconsin where Judge Crabb, a democrat appointee, denied Epic System’s motion to compel arbitration. The 7^th Circuit panel comprised of Judges Wood (a democrat appointee), Rovner (a republican appointee), and Blakey (a democrat appointee), voted to affirm the Judge Crabb’s decision. This worked out to be a partisan relationship where the majority appointing party of the appeals court judges accorded with the appointing party of the district court’s judge and the district court’s ruling was affirmed. The Supreme Court then reversed the appeals court’s decision with a 5 to 4 vote along partisan lines. Justice Gorsuch authored the Court’s majority opinion and was joined by Justices Alito, Thomas, Kennedy, and Roberts, all republican appointees. The dissenters, all democrat appointees included Justices Ginsburg, Breyer, Sotomayor, and Kagan.  Since a majority of republican appointees in the Supreme Court reversed the decision of a majority of democrat appointees in the appeals court, this relationship was coded as partisan as well. In effect the three decisions in the Epic Systems chain led to two partisan relationships out of two possible relationships.  This reasoning was then applied to all of the Court’s decisions for the term.  If partisanship really is not a factor in decision making then we should have random outcomes or around a 50% total percentage of partisanship. The role of partisanship in the decisions will be more obvious, however, if the partisanship level of decision making lands perceptibly above this 50% marker.

There are several caveats to this analysis at the outset. First, there were some discretionary decisions in terms of how to code case outcomes. The Court’s decision in Gill v. Whitford was to vacate the lower court’s decision based on standing grounds. Decisions vacating a lower court result are often in favor of the petitioner as was the case in Byrd v. U.S. which was argued later in the term. For the purpose of this analysis, the result in Gill was coded as null at the Supreme Court level because it was premised on justiciability (standing in that instance) rather than on the case’s merits.

Across the Court’s entire first sitting, 11 of the 15 coded decisions or about 73% accorded with partisan alignment. Along with Epic Systems, these included both courts’ decisions in Ernst & Young LLP v. Morris, the Supreme Court’s decision in National Labor Relations Board v. Murphy Oil USA (the appeals court in the case granted and denied the petition in part), the Supreme Court’s decision in Sessions v. Dimaya (the trial court level was handled by a federal agency in that case), both decisions in Jennings v. Rodriguez (since the appellate court mainly affirmed the district court’s decision this was coded as partisanship alignment), and both decisions in District of Columbia v. Wesby. The Supreme Court’s decision was coded for partisan alignment in Class v. U.S. yet the lower court’s was not as a majority republican panel affirmed the decision of a democrat appointed district court judge. No outcome was coded in Hamer v. Neighborhood Housing Services of Chicago since the appeals court dismissed the case and the Supreme Court remanded the case back to the appeals court based on the lower appeals court’s error (the conclusion of Ginsburg’s majority opinion language in Hamer includes in this regard, “We note…that our decision does not reach issues raised by Hamer, but left unaddressed by the Court of Appeals.)  Finally  the outcomes were coded as non-partisan both in Jesner v. Arab Bank and in the Supreme Court for National Association of Manufacturers v. Department of Defense (NAM).  The trial court level of NAM involved litigation in multiple districts and so a binary label on the appeals court decision would not capture the effects of the decision in the case.

In its simplest form the partisanship measure of the decisions for the October 2017 sitting looks like:

• Ernst & Young LLP v. Morris: 2/2
• Epic Systems Corp. v. Lewis: 2/2
• National Labor Relations Board v. Murphy Oil USA: 1/1
• Sessions v. Dimaya: 1/1
• Gill v. Whitford: NA
• Jennings v. Rodriguez: 2/2
• District of Columbia v. Wesby: 2/2
• Class v. U.S.: 1/2
• Hamer v. Neighborhood Housing Services of Chicago: NA
• Jesner v. Arab Bank, PLC: 0/2
• National Association of Manufacturers v. Department of Defense: 0/1

The trajectory of the Jesner case is an acute counterexample to the theory that partisanship matters in federal judicial decision making. In Jesner a Second Circuit panel composed of a majority of democrat appointees affirmed the decision of Judge Cogan, a republican appointee to the Eastern District of New York.  The panel’s decision was then affirmed by not only a majority of republican appointed justices but by a 5 to 4 margin with only republican appointees in the majority and only democrat appointees in dissent.  Even as this example shows the grip of partisanship is not inexorable, that is not to diminish its implications in decision outcomes.

When extrapolated across the entire 2017 these parallel party of appointing president similarities and alternative party differences, marked by rulings in the same or different direction in the same cases, are the norm. The figure below shows the percentage of parallel and alternative direction rulings (depending on the partisanship of the appointing president) in each case pipeline by Supreme Court oral argument sitting.

Each sitting aside from the fourth had at least half of its decisions marked by partisan differences. The first three sittings each had over 70% partisan decision making (73.3%, 72.7%, and 75% respectively). The average percentage across sittings does not equate to the total mean value since the number of coded arguments per sitting differs.

While the partisan markers in the above analysis are not overwhelming in each instance, we should not expect them to be so.  There is much room for agreement between judges with alternate beliefs on many issues. The differences shown in the statistics above are well beyond random happenstance though.

This does not mean that judges are actively thinking that their decisions should accord with other judges appointed by presidents from the same party, and their decisions should differ from those appointed by presidents from the alternate party. Moreover, partisanship is likely not even the best measure of difference between federal judges.  Preference based measures tend to perform much more strongly than the appointing president measure when trying to predict judicial votes. The fact is though that presidents seek likeminded judges for the federal judiciary, and these likeminded judges often have different priors than judges nominated by presidents who also have different priors. [If you are interested in learning more about how these measures apply to members of the federal judiciary you might enjoy Epstein, Landes, and Posner’s work The Behavior of Federal Judges.]

We still should not reduce a judge’s role to mere voting behavior. Appellate judges write opinions that are not necessarily binary but often reflect a spectrum of views that cannot be easily reduced “affirm” or “reverse.” When looking at voting outcomes, however, partisanship, at least in terms of appointing president, helps to dictate the decisions of federal judges in complex cases moving through the federal judicial hierarchy.

Chief Justice Roberts’ statement defending an independent judiciary is accurate to a point. While an “Obama judge” might not differ much from a “Clinton judge,” he or she will often have differing beliefs from a “Trump” or “Bush judge.” Judges are real people and as people they differ in beliefs. Any statement to the contrary elides significant research to the contrary (for example). While reality does not permit us to run the counterfactual in the same instance of a republican appointed judge deciding on President Trump’s asylum policy, Judge Tigar’s party affiliation might have, at least in part, helped dictate his view in the matter.

Idealistically we, like the Federalists, might hope and strive for pure and neutral judges who decide according to the rule of law and cast away person preferences.  But this is just that – somewhat fantastical.  Instead of dealing with a fiction, we might be better off accounting for differences in judges that sometimes boil down to divergences pinpointed by presidents of opposing parties, and acknowledge that those differences have implications for federal judges’ decisions.

You can get access to the entire term’s set of case pipelines on the data page at Optimized Legal (free login required).

On Twitter: @AdamSFeldman