In new research, Alma Cohen finds that the political affiliations of Circuit Court judges influence decisions in a much wider variety of cases than previously thought.

Because different judges are supposed to implement the same legal rules, judicial decisions are generally expected not to depend on the political affiliations or ideological leanings of the deciding judges. However, it is widely believed that in the United States Supreme Court, which often examines ideologically controversial cases, there are systematic differences between the decisions of “Democratic” and “Republican” justices—that is, justices nominated by Democratic and Republican presidents, respectively. Similarly, empirical studies have found that Democratic and Republican judges serving on federal Circuit Court of Appeals systematically differ in their decisions on “ideologically controversial issues,” such as abortion, affirmative action, capital punishment, and sex discrimination.

However, many scholars studying the Circuit Courts believe that “outside of such [ideologically salient] domains, [Republican and Democratic judges] are far less likely to differ.” In a new study, I ask if Republican and Democratic judges are in fact similar in their rulings in the vast number of cases that are not considered by the literature as ideologically controversial and find that the political affiliations of judges does help to predict decisions in a large universe of cases that are not on ideologically controversial issues.

Democratic judges tend to favor the weaker party

To study this question, I assembled a large dataset comprising approximately 670,000 circuit court cases spanning a 35-year period starting from 1985. This dataset is much larger than the datasets of ideologically salient cases with published opinions on which much of the empirical literature has thus far examined.

To discern between the views of Republican and Democratic judges, I first focused on cases whose characteristics suggest that the parties might be of “unequal power.” For these cases, I hypothesize the following Pro-weak hypothesis: In litigation between parties of seemingly unequal power, the higher the number of Democratic judges on the panel (there are usually three at the Circuit level), the greater the tendency for the panel to side with the seemingly weaker party.

Specifically, I review the following category of cases: (i) non-institutional private parties in civil litigation against institutional private parties; (ii) private parties in civil litigation against the U.S. government; (iii) convicted offenders in criminal appeals litigated with the U.S. government; (iv) immigrants in litigation against immigration agencies; (v) prisoners serving a sentence in litigation with the government and prisons and (vi) petitioners in habeas corpus and other petitions against public officials for which the circuit courts have original jurisdiction. These sets of cases represent together over 80% of all circuit court decisions.

There are several reasons for advancing this pro-weak hypothesis. First, Democratic judges could be more likely to hold liberal legal positions, which in many cases correlate with favoring the weaker party. Additionally, Democratic judges might have a greater sympathy for the weak party. Furthermore, Democratic judges may be more concerned about the disadvantaged position of the weaker party in the preceding stages of the case.

I finds that, in each of the above categories of cases with parties of unequal power, having a higher number of Democratic judges on the panel is associated with higher likelihood of an outcome that is favorable to the weaker party.

The identified association is not merely statistically significant but also meaningful in magnitude. To illustrate, for the approximately 550,000 cases with parties that are seemingly of unequal power, switching from an all-Republican panel to an all-Democratic panel is associated with an increase of 55% in the baseline likelihood of an outcome favoring the weaker party. Thus, the likelihood of a pro-weak outcome significantly depends on the political affiliations of the judges randomly assigned to the case, and thus on the “luck of the draw.”

To examine whether my results are driven by cases that are ideologically salient, I used the methodology of Cass R. Sunstein, David Schkade and Lisa Michelle Ellman to identify ideologically contested cases if they involve any of the fourteen categories specified by their study to be saliently ideological. I find that the patterns I identify are not confined, or even driven by, these ideologically salient cases. Rather, it is also widespread in a majority of the cases which were not classified as ideologically using the methodology of Sunstein et al. 

The results should not be interpreted to suggest that the association between political affiliations and rulings is present in all non-ideological cases. Instead, they suggest that the vast number of cases that are not notably ideologically salient are nonetheless “ideological” in the sense that Republican and Democratic judges systematically differ in how they approach them.

Democratic judges are also more willing to intervene in lower-court decisions

After analyzing cases with seemingly unequal parties, I turn to cases that do not appear to have such inequality. For these cases, I explore the Pro-Intervention (or Less Deference) hypothesis. This hypothesis posits that in litigation where the parties appear to have equal power, a higher number of Democratic judges on the panel increases the likelihood of the panel intervening in the district court’s decision. This hypothesis derives from the possibility that Democratic and Republican judges may value the benefits of deference differently due to Democratic judges placing less weight on the resource-saving efficiency gains from such deference and more weight on reaching the right outcome in each individual case.

To test this hypothesis, I analyzed approximately 80,000 civil appeals where it was not evident that one of the parties was stronger than the other. My findings from this extensive sample of cases in which the parties are presumed to have equal power are consistent with this hypothesis.

Lone Republicans have more impact on panel decisions than lone Democrats

Lastly, I investigated whether Democratic and Republican judges are similarly influenced by their peers. Given that circuit court judges typically make decisions in panels of three, it is plausible that when judges make decisions, they are shaped not only by their individual preferences but also by the influence of their colleagues. Having more judges from the same party on the panel might amplify a judge’s individual preference, whereas having on the panel members from the opposing party might moderate these preferences.

To assess this question, I compared the impact of removing a lone Democrat (shifting from two Republicans and one Democrat to an all-Republican panel) to that of removing a lone Republican (moving from two Democrats and one Republican to an all Democrat panel). The results indicate a disparity, with the removal of a lone Republican having a more pronounced effect than the removal of a lone Democrat.

This disparity may be attributed to Democratic judges being more susceptible to amplification or group polarization effects in a “single-party” panel. Additionally, Democratic judges may be more open than Republicans to being persuaded by a lone judge of the opposing party, or more inclined to accommodate a lone judge from the opposite party in order to avoid conflict, prevent a dissenting opinion, and secure a unanimous decision and collegial outcome.

In sum, my analysis has shown that the judges’ political affiliations, inferred from the party of the appointing president, can be used as a predictive tool for decision outcomes in 92% of the circuit court decisions studied. This association is far more widespread than previously documented in prior research.

Going forward

I would like to stress that while my study identifies systematic differences between the decisions of Democratic and Republican judges, it does not attempt to determine or evaluate which judges make “better” decisions. Whereas I show that Democratic judges are more likely to favor the weaker party in cases with parties of seemingly unequal power, I do not attempt to assess whether Democratic judges are overprotective, or Republican judges are insufficiently protective of weaker parties. And whereas I find that Republican judges are more likely than Democratic judges to defer to the district court decision, I do not attempt to assess what level of deference is better. My goal has been to map the widespread systematic differences between the decisions of Republican and Democratic judges.

Furthermore, it is worth stressing that my results do not imply that political affiliations fully determine judicial decisions. Legal factors undoubtedly play a substantial role in shaping outcomes. However, whereas political affiliations do not fully determine or predict judicial outcomes, my analysis shows that they can definitely help to predict such outcomes. My study uses very simple, coarse, and easily observable characteristics of cases to predict how panels of different political composition would decide cases. Future research could build on this foundation, for example, using less noisy proxies for inequality of power than I did. More work should be done to fully map the patterns of association between political affiliations and outcomes in circuit court decisions.

Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.