Panel Effects and Precedent

The vast majority of circuit law is established by three-judge panels. Panel assignments rotate on a regular basis, resulting in each judge working with various combinations of his circuit colleagues. Sometimes a judge will sit with two like-minded colleagues, while other cases have to be resolved working with colleagues who have quite different ideological preferences. This institutional feature creates interesting variation in the output from different panel configurations. Most prominent is the robust finding in the literature that even a single judge can have an impact on case outcomes (see, e.g., Cross and Tiller Reference Cross1998; Sunstein Reference Sunstein2006; Boyd et al. Reference Boyd, Epstein and Martin2010; Kastellec Reference Kastellec2013; Beim, Hirsch, and Kastellec Reference Beim, Hirsch and Kastellec2016). Panels resolve cases by majority rule, yet empirical results consistently reveal patterns at odds with what the median voter theorem would predict (Kastellec Reference Kastellec2007). The primary explanations for these findings are strategic dissent avoidance to circumvent external review (Cross and Tiller Reference Cross1998; Kastellec Reference Kastellec2011; Beim et al. Reference Beim, Hirsch and Kastellec2016), strategic dissent avoidance to reduce workload and comply with a norm of consensus (Posner Reference Posner2008; Epstein, Landes, and Posner Reference Epstein, Landes and Posner2011; Fischman Reference Fischman2015), and internal deliberation leading to incorporation of a broader range of information (Kim Reference Kim2009; Spitzer and Talley Reference Spitzer2013).

Although the bulk of the literature on panel effects focuses on cases outcomes, there is some work that sheds light on the impact that panel configuration has on aspects of the opinion-drafting process. Haire et al. Reference Haire(2013) demonstrate that the presence of women and racial minorities on a panel can increase the number of legal issues discussed in the final opinion, although a panel majority is necessary to generate a significant effect. There is also empirical evidence that panels with members who have more divergent ideological preferences take longer to produce a final opinion (Bowie et al. Reference Bowie2014, 110). Such delay suggests that increased ideological diversity leads to increased negotiation over opinion content. These studies indicate that panel effects can manifest in opinion writing as well as in voting.

Exploring the role of panel effects on opinion drafting is important because of the key role opinion content plays in legal development. As Songer, Sheehan, and Haire (Reference Songer2000) note, “most judges consider opinion writing to be the most important part of their work because the court’s decision will make new law or clarify existing law for those in the legal community” (11). The details of a published opinion can have important implications for a range of future actors. Many people who will never end up in court may change their behavior in anticipation of how the opinion indicates a future conflict would be resolved. More directly relevant here, a range of future litigants will bring claims to court that may be controlled by that precedent. The effect of the precedent in such cases is not static. Studies have shown that how a precedent is used changes its impact over time (Hansford and Spriggs Reference Hansford2006; Westerland et al. Reference Westerland2010). When circuit courts treat Supreme Court precedents more frequently, either positively or negatively, that leads to an increase in the same type of treatment in later circuit cases (Westerland et al. Reference Westerland2010). Furthermore, when the Supreme Court narrows the scope of its own precedents through negative treatment, that also increases the probability that it will be narrowed even more in subsequent cases (Hansford and Spriggs Reference Hansford2006). These results, taken together, suggest that how a circuit uses its own precedents will influence the continued scope of those precedents. Specifically, I hypothesize that a precedent that is negatively treated (or favorably cited) more frequently in a given year will be negatively treated (or favorably cited) more frequently in the following year.

Decisions that influence the subsequent use of a precedent are important to study because the use of a precedent reflects the impact it has on legal doctrine. The development of law is nuanced, complicated, and difficult to measure. But citations and treatment provide an opportunity to measure this key concept in a way that is useful, if not perfect. Opinions that are favorably cited more often have a larger influence on legal policy than opinions that are cited less often (or not at all; Landes, Lessig, and Solimine Reference Landes1998). Admittedly, not all “favorable” (i.e., nonnegative) citations carry the same level of significance. Brief inclusion in a string citation indicates less influence than a more in-depth discussion of a case. However, in spite of such limitations, favorable citations can provide valuable insight (e.g., Klein and Morrisroe Reference Klein1999).Footnote ³ Academics have long viewed citations to their own work as a useful measure of scholarly influence (Margolis Reference Margolis1967). Arguably, such impact is even more consequential when the documents at issue constitute binding precedent. As a result, exploring the role panel effects play in decisions about the use of binding precedents ultimately sheds light on the dynamic process of legal development.

Authors and the Impact of Other Panel Members

The important role a counterjudge can play in who wins and loses suggests that similar panel effects might extend to how the majority opinion is crafted to influence the development of circuit law. Unlike voting, crafting an opinion is not an activity that divides power equally among the panel members. The responsibility of drafting an opinion to which the other panel members respond gives the author the potential to wield disproportionate power. Moreover, practical considerations and institutional norms dictate that the three members of a panel share opinion-writing duties equally across the subset of cases entrusted to the panel. As a result, over time almost all circuit judges will not only serve on both split and unified panels and find themselves in both the panel minority and majority, they will also be entrusted with authoring opinions under each possible panel configuration.

The author’s disproportionate power is the reason scholars have turned their attention to studying the process of opinion assignment. In the Supreme Court context a line of research has demonstrated that chief justices exercise their opinion-assigning power in an ideological and strategic fashion, even changing their vote to the majority in some instances in order to avoid ceding that power (Brenner and Spaeth Reference Brenner1986; Maltzman and Wahlbeck Reference Maltzman1996; Maltzman, Spriggs, and Wahlbeck Reference Maltzman2000). The power of assignment is valuable because the opinion author possesses a first-mover advantage in framing the scope and content of the legal rule (Murphy Reference Murphy1964; Maltzman et al. Reference Maltzman2000; Lax Reference Lax2007). We currently know less about the politics of opinion assignment at the circuit level, but what we know suggests assignment is done in a more collaborative fashion (Cheng Reference Cheng2008; Farhang, Kastellec, and Wawro Reference Farhang2015). Regardless of how opinions are assigned, circuit court authors have considerable power since they frame the legal rule set forth in the opinion and face considerably less input from their colleagues than occurs at the Supreme Court (Choi and Gulati Reference Choi2007). Rather than explore why opinions are assigned to particular judges, this article focuses on the question of whether panel configuration influences how the assigned author carries out that task.

While circuit judges possess a clear opportunity to exercise power when they draft an opinion, there are also very real limitations on that power. Most obvious is that the case outcome is generally determined in advance. Although the deliberative process is secret, we do know that it typically involves registering preliminary votes. Only after a tentative determination about the outcome of the case is one judge assigned to draft the majority opinion (Bowie et al. Reference Bowie2014, 62–65). Consequently, a major limitation on the parameters of how an author writes an opinion is that the outcome has been determined in advance. Any change to the case outcome after the majority opinion draft is circulated is unusual. During an interview, one circuit judge estimated that votes change after the preliminary vote at the conference “in maybe 8 or 9 percent of the cases” (99). Another obvious limitation is that a judge is constrained to write about the facts of a case and the legal issues they raise. The following discussion focuses on the discretion an author can wield within these constraints.

Even in the typical case in which votes on the outcome remain constant from the conference to the release of the finalized majority opinion, authors still exercise discretion about how to draft an opinion. As Kim Reference Kim(2009, 1343) notes, “the reasons that they [judges] give to justify their decisions are critical, for it is the content of opinions rather than the simple declaration of a winner that shapes the development of the law.” These reasons are primarily crafted by the opinion author. Although the other two judges on a panel retain the power to ask for changes or withdraw their support of a draft opinion, there is reason to expect that circuit judges give a fairly high level of deference to the author, at least compared to the Supreme Court (Choi and Gulati Reference Choi2007). The sheer magnitude of circuit court caseloads makes efficiency a necessity rather than a luxury. Faced with mandatory jurisdiction, increasing caseloads, and frequent lengthy vacancies in existing circuit court seats, circuit judges have limited resources to devote to the consideration of each case (Cheng Reference Cheng2008; Wood Reference Wood2012; Farhang et al. Reference Farhang2015). Furthermore, when an author includes a favorable citation to a binding precedent, legal doctrine may shield that decision from criticism. Beim and Kastellec Reference Beim(2014) note that counterjudges who are aligned with circuit law are in a stronger negotiating position. Similarly, authors, even counterjudge authors, who are aligned with a binding precedent can rely on stare decisis to support their ideologically preferred course of action.

Although concerns about efficiency, effort aversion, and legal doctrine may result in deference to the author, they may also result in deference by the author to the preferences of her panel colleagues. A central element in the various explanations for panel effects is avoiding separate opinions. According to whistleblower theory, dissenting opinions are avoided to reduce the probability of external review (Cross and Tiller Reference Cross1998; Kastellec Reference Kastellec2011; Beim et al. Reference Beim, Hirsch and Kastellec2016). Deliberative explanations incorporate judges’ general unwillingness to openly disagree with colleagues as one of the reasons they are open to being persuaded (rather than writing separately; Haire et al. Reference Haire2013; Spitzer and Talley Reference Spitzer2013). Finally, internal strategic explanations emphasize the extra time and effort required by separate opinions (Posner Reference Posner2008; Epstein et al. Reference Epstein, Landes and Posner2011; Fischman Reference Fischman2015). The shared element is the various benefits judges enjoy when they vote in a manner that prevents separate opinions. Panel effects on opinion writing may similarly be generated by authors working to avoid separate opinions.

There are two ways authors may reduce the probability of separate opinions. First, authors may accommodate revisions requested by other panel members after the draft majority opinion is circulated. Second, the author may anticipate the preferences of the other panel members and initially draft an opinion that is designed to satisfy their concerns. Deliberation about the content of an opinion would manifest as the first type of dynamic. Nonauthor panel members who are ideologically diverse may bring a wider range of information to the case and suggest the inclusion of additional circuit precedents or make a persuasive case for a different approach to discussing cited circuit law. Evidence from interviews of circuit judges indicates that quite a bit of bargaining over the content of published opinions does take place, in spite of their busy caseloads (Bowie et al. Reference Bowie2014, 93–99). Therefore, there is plenty of opportunity for nonauthors to weigh in about various nuances of an opinion, including the use of precedent. The fact that panels take a longer time to issue the final opinion when the panel is more ideologically diverse suggests that differing views on a panel may tend to generate more back-and-forth over the opinion content (110). Since precedents from the same circuit are binding as a matter of legal doctrine, nonauthors have a particularly strong doctrinal foundation for asking an author to incorporate such precedents (Beim and Kastellec Reference Beim2014; Tiller Reference Tiller2015).

In addition to the iterative circulation of opinion drafts, circuit judges discuss the fact that their initial drafts are constructed with the other panel members in mind (Bowie et al. Reference Bowie2014, 97–98). One interviewed judge reported that “if you are assigned to write the court opinion you will always take into account what the preferences of your colleagues are” (97). The motivation for such anticipation can be framed in strategic terms. Delays in getting one opinion finalized do not prevent an author’s writing obligations in other cases from piling up. The swiftest way to dispatch with an authorship assignment is to circulate an initial draft that satisfies both panel colleagues with minimal revisions. When a draft raises concerns for one or both of the nonauthors, more effort is required by the author whether the concerns are accommodated or whether they lead to the other judge writing a separate opinion. The presence of a separate opinion both delays the resolution of the case in general and creates additional work for the author of the majority opinion because it is typical to revise the majority opinion to address the arguments raised in the separate opinion (Epstein, Landes, and Posner Reference Epstein, Landes and Posner2013; Haire et al. Reference Haire2013; Bowie et al. Reference Bowie2014). Authors can minimize such complications by anticipating the preferences of their panel colleagues.

An author can craft an opinion to frame a legal rule narrowly or broadly. That framing can influence how many future cases will be controlled by the same reasoning. Yet gaining theoretical and empirical traction on such nuances poses serious challenges. These are mitigated by the central role of precedent in legal reasoning. In the abstract a judge’s task is to determine which previous cases are similar in important respects and which are not. In reality, the panoply of precedents available to apply in any given case potentially gives authors the opportunity to cherry-pick those which they prefer and ignore (or explicitly reject) other precedents (Cross et al. Reference Cross2010; Niblett Reference Niblett2010). According to the doctrine of stare decisis, all published circuit court opinions are binding within their own circuit (Barnett Reference Barnett2002; Lee Reference Lee2003). However, empirical research shows that judges’ own ideologies influence when and how they comply with stare decisis (Spriggs and Hansford Reference Spriggs and Hansford2002; Choi and Gulati Reference Choi2007; Hume Reference Hume2009; Niblett and Yoon Reference Niblett2015).

An author seeking to shape circuit law in accordance with his own preferences will seek to favorably cite ideologically proximate precedents and ignore or negatively treat ideologically distant precedents. If panel effects constrain how an opinion is crafted, the effect of the author’s ideology on citation and treatment decisions will be dampened when the panel is split compared to when the panel is unified. Such ideological dampening should be greatest when the author is the counterjudge because such a judge will be concerned about shaping an opinion amenable to two ideologically distant colleagues. A majority judge writing for a split panel will only have to accommodate the preferences of one such colleague. In short, if the way an author crafts an opinion is influenced by panel composition, each additional panel member from a party different from the author’s should increasingly dampen the impact of the author’s ideology. A judge writing to convince an ideologically distant panel member (or two) may not have the luxury of narrowing or ignoring those precedents located further away from herself.

Panel effects on opinion crafting could take other forms in addition to ideological dampening. An author’s deference to his panel colleagues may emerge on a personal level. Even after accounting for other factors, an author may be more deferential toward the precedents written by one of the other panel members. Evidence that an author is less likely to negatively treat, or more likely to favorably cite, such a precedent would indicate that the identity of the nonauthoring judges on a panel can influence the content of the opinion. Like ideological dampening, such a pattern may emerge as a result of either accommodation or anticipation. First, a nonauthor might be more likely to ask for inclusion of a precedent she wrote. Second, the author might initially seek out precedents written by nonauthors, anticipating that they will both have more information about such precedents and care more about whether they are discussed. Similarly, an author may choose to ignore rather than negatively treat an ideologically distant precedent written by a fellow panel member. As a result, authors may tend to produce opinions that are disproportionately deferential toward the precedents written by their fellow panel members.

Analyzing the impact of nonauthors on opinion content requires controlling for other factors that influence decisions about precedent that may also be correlated with either author ideology or panel type. As discussed above, authors seek to maximize efficiency when crafting an opinion. One way to increase efficiency is to rely on precedents with which the author is already familiar. A judge will be most familiar with those precedents she actually wrote, making favorable citation more likely. Another goal is the pursuit of reputational benefits, which makes negative treatment of a precedent written by the author less likely. Finally, efficiency concerns may result in busier judges in circuits with higher caseloads ignoring more precedents because of the increased time pressures they face.

A number of features of a precedent reflect its strength in terms of legal doctrine. Judges may seek to employ stronger legal precedents to pursue a variety of goals: building professional reputation, increasing persuasiveness, preserving legitimacy of the legal system, disposing of cases efficiently, or fulfilling their perception of the appropriate judicial role (Hansford and Spriggs Reference Hansford2006; Posner Reference Posner2011). While I do not disentangle these goals, I do control for a variety of features of a precedent that legal doctrine suggests are applicable. Central to the process of analogical reasoning is the proposition that judges should apply relevant precedents (Schauer Reference Schauer1987; Aldisert Reference Aldisert1989). Precedents that are more similar to the panel opinion should also be more legally relevant (Hinkle Reference Hinkle2015, Reference Hinkle2016).

In addition to relevance, several other characteristics of a precedent indicate its doctrinal strength. Precedents accompanied by a dissenting opinion provide a less firm foundation since an alternate legal viewpoint is explicitly recorded (Johnson Reference Johnson1979; Posner Reference Posner2008). Precedents resolved per curiam reflect the judgment that the case is not particularly important (Hume Reference Hume2009; Hinkle Reference Hinkle2015). Conversely, en banc precedents are unusually strong since circuits only review the most important cases en banc, and such rulings benefit from the combined wisdom of a larger number of judges (Hume Reference Hume2009). A longer precedent may indicate more thorough and detailed legal analysis, or it may simply cover a wider array of legal topics. Whether precedent length reflects quality or complexity (and, therefore, applicability to a wider range of subsequent cases) it should be correlated with a higher probability of favorable citation (Black and Spriggs Reference Black2008; Hinkle Reference Hinkle2016).

Over time the strength of each precedent changes as well (Gerhardt Reference Gerhardt2008; Black and Spriggs Reference Black2013). This is due to the simple passage of time as well as the impact of subsequent cases that cite or treat a precedent. As a result, it is important to consider the age of a precedent and its legal strength at the time it is considered for inclusion in an opinion. The vitality of a precedent summarizes how its scope has been expanded or narrowed by subtracting the number of negative treatments of a precedent from the number of times it has been positively treated (Hansford and Spriggs Reference Hansford2006). The overall number of times a precedent has been cited and treated also reflects its strength (Hansford and Spriggs Reference Hansford2006). These dynamic variables help account for path dependency that can occur in citation practices. Cases that are cited more frequently tend to pick up momentum in terms of future cites not only because of efficiency concerns but also because each repeated use of a precedent contributes to its strength as a legal precedent. In fact, these patterns are precisely why it is important to examine whether panel effects affect how precedents are used.

Other features of the context within which the author makes citation decisions may also be relevant. Both the length of the opinion itself and the number of precedents available to choose from are important. A longer opinion can discuss more cases than a shorter opinion. Conversely, a larger available choice set gives the author more choices, which makes it less likely overall that any given precedent will be addressed.