What influence does Supreme Court activity have on the issues that the lower federal courts address? While scholars have sought to understand whether lower courts comply with the edicts handed down by the Court on a particular issue (e.g., Reference Cameron, Segal and SongerCameron, Segal, & Songer 2000; Reference KastellecKastellec 2007; Reference LaxLax 2003; Lindquist, Haire, & Reference Lindquist, Haire and SongerSonger 2007), research has rarely examined how the Court's decisions shape the very issues addressed in the lower federal courts. If the Court's decisions influence the agenda of the federal courts, then the Court shapes both policy and lower court opportunities for compliance with the Court's preferences on that policy.
On this question, two separate perspectives provide divergent expectations. The conventional legal perspective suggests that the Court decides cases in order to resolve unsettled areas of law. In this view, Supreme Court decisions lead to fewer subsequent disputes within an issue area after the Court addresses that issue (Reference Casper and PosnerCasper & Posner 1974; Reference FriedmanFriedman 1967). Alternatively, recent research on interest group influence on the judicial agenda proposes a different view. This interest group perspective suggests the Supreme Court actually encourages additional litigation through signals to litigants (Reference BairdBaird 2004, Reference Baird2007; Reference Baird and JacobiBaird & Jacobi 2009b). Signals from the Court are received by litigants, who bring or support cases that percolate up through the judicial hierarchy, eventually leading to temporary increases in attention to issue areas the Court has recently addressed. Thus, these two perspectives provide divergent expectations for what happens in the federal courts after the Supreme Court acts.
The goal of this article is to parse these theories of the Supreme Court's agenda control and to determine what happens to the lower federal courts’ agendas after Supreme Court decisions. Utilizing filings on case terminations from the Administrative Office (AO) of the Federal Courts, the published opinions of federal courts of appeals and district courts, and supervised machine learning methods for text classification, I present a comprehensive and extended portrait of issue attention in the federal courts, as well as the Supreme Court's role in determining it. In accord with the legal perspective, I find evidence in both trial and appellate courts that Supreme Court attention to policy areas subsequently leads to fewer cases being heard and decided in those policy areas in the lower courts. Yet I also find evidence of additional interest group attention, and additional published opinions, in lower federal courts in issue areas after the Supreme Court addresses that issue.
Taken together, these findings indicate that the Supreme Court exerts considerable control over the agenda of the federal courts, a critical power in the formation of public policy. In so finding, this research has important implications for our understanding of issue attention relationships within the judicial hierarchy, the broad dynamics of the federal judicial agenda, and the strategies of litigants and interest groups. The Court's attention shifts the very participation of certain actors seeking to influence public policy in the federal courts, as issue areas go from being characterized by broad-based litigation to being characterized by less litigation, but more sophisticated participants. While potential litigants—concerned generally with the outcome of their individual cases—gain clarity from Supreme Court decisions on their probability of success, and are therefore less likely to require the courts (Reference Priest and KleinPriest & Klein 1984), interest groups—concerned primarily with public policy—respond to Supreme Court decisions on an issue by devoting additional attention to that same issue. At the juncture of these dynamics, Supreme Court attention shifts the issue attention of the federal courts, leading to important and understudied implications for studies of hierarchical relationships. The Supreme Court, in deciding to devote attention to particular issue areas, systematically alters the attention the federal courts devote to that issue, and thus the influence the judiciary has on that issue, in subsequent years.
Judicial Compliance and the Importance of the Agenda
A rich history of research on the judiciary has sought to understand how, and how effectively, the Supreme Court exercises hierarchical control over lower federal courts. Much of this research aims to determine the extent to which the Supreme Court can ensure lower courts comply with the Court's preferences. For the most part, scholars have found lower court judges tend toward relatively faithful adherence to the Court's preferences (Reference Songer, Segal and CameronSonger, Segal, & Cameron 1994; Reference WesterlandWesterland et al. 2010). Accordingly, other research has teased out the methods by which this hierarchical control operates, including signals as to cases which necessitate review (Reference Cameron, Segal and SongerCameron, Segal, & Songer 2000), panel influences (Reference KastellecKastellec 2007) and litigant selection processes (Reference Cameron, Kornhauser, Bond, Flemming and RogersCameron & Kornhauser 2006).
The unit of analysis in these studies is typically the vote or outcome of individual cases, with the goal to understand hierarchical control of judicial decisionmaking. The research thus addresses the decision-making stage, but not the prior step in which the cases actually arrive in lower courts. The ability to manipulate this step—to determine what is on the agenda of the lower federal courts—is crucial to understanding the decision-making process (Reference Bachrach and BaratzBachrach & Baratz 1962). Having an issue discussed is the first, and most integral, step in effecting policy change (Reference PacellePacelle 1991; Reference SchattschneiderSchattschneider 1960).
Yet while research has provided a number of avenues for presidential and Congressional influence on the issue attention of courts (e.g., Reference Johnson and CanonJohnson & Canon 1984; Reference SalokarSalokar 1992), we do not know whether and how the Supreme Court influences what lower federal courts discuss and decide. Yet history suggests influence does exist. As but one of many potential examples, in Baker v. Carr Footnote 1 the court determined that legislative redistricting was within the jurisdiction of the federal courts. Forty years later, nearly every state has been involved in some form of federal redistricting litigation (Federal Court Involvement in Redistricting Litigation 2001). In deciding the case, the Court directly encouraged future judicial attention to the issue of redistricting.
Beyond this one case, how does the activity of the Supreme Court systematically impact the agenda of the federal courts? While it is true that any one Supreme Court decision may positively or negatively influence subsequent attention in lower federal courts, my goal is to determine the general pattern of issue attention influence. Because the federal courts do not have discretionary jurisdiction, this is ultimately a question of how litigants respond to Supreme Court decisions. The agenda of the federal courts, and the policy-making power of these courts, is dependent on the flow of litigation (Reference EppEpp 1998), and the impact of litigants in shaping the agenda has accordingly been the subject of extensive study (Reference BairdBaird 2004, Reference Baird2007; Reference EppEpp 1998; Reference PacellePacelle 1991; Reference PetersPeters 2007). Two separate perspectives prevalent in prior literature offer different predictions for how litigants respond.
The Legal Perspective
Sitting at the top of the judicial hierarchy in the United States, the Supreme Court is the final arbiter in the interpretation of federal law. With limited resources, the Court can only decide a small number of cases every year despite an overwhelming volume of petitions for review. From the legal perspective, the Court chooses to hear and decide cases so as to “… further the ‘legal’ goals of consistency and certainty in the law” (Reference BaumBaum 1977: 14). By deciding the cases, the Court's decisions establish or amend legal rules that are ostensibly intended to guide the decisionmaking of lower courts and other institutions (Reference FriedmanFriedman 1967), and potentially stem a tide of litigation on that issue (Reference WahlbeckWahlbeck 1998). Thus, by deciding cases, the Court “minimizes the risk of further litigation and maximizes the extent to which other private or public agencies can apply the rule, thus taking pressure for decision away from the courts” (Reference FriedmanFriedman 1967: 815).
Note that the structure of the federal court system explicitly encourages settling of the law in this way. What little guidance exists for how the Court actually chooses cases lies in Rule 10 of the Supreme Court Rules, which specifically identifies resolving areas of conflictual interpretation as a criteria for Supreme Court review. Even more importantly, Rule 10 states that consideration should be given when a lower court “… has decided an important question of federal law that has not been, but should be, settled by this Court or has decided an important federal question in a way that conflicts with relevant decisions of this Court.” In short, the express guidance provided by Rule 10 indicates the priority of settling law, whether due to conflict in interpretation among courts or because the Court has yet to address an important issue. By design, then, the Supreme Court is to choose cases to settle law in the lower courts.
Accordingly, justices generally focus on identifying and selecting cases with wide-ranging importance and contradictory resolutions in lower courts. Research on specific factors that influence the likelihood of review (see, e.g., Reference Brenner and KrolBrenner & Krol 1989; Reference Epstein and KnightEpstein & Knight 1998; Reference PerryPerry 1991; Reference ProvineProvine 1980; Reference Segal and SpaethSegal & Spaeth 1993; Reference Tanenhaus and SchubertTanenhaus et al. 1963; Reference Teger and KosinskiTeger & Kosinski 1980; Reference UlmerUlmer 1984) has borne out the importance of these two factors. Moreover, the Court uses these cues and others to determine cases which provide the best vehicles for ensuring lower court compliance with the Court's preferences (Reference Cameron, Segal and SongerCameron, Segal, & Songer 2000; Songer, Segal, & Reference Songer, Segal and CameronCameron 1994).
In deciding the case, the Court lays out legal rules that are intended to guide future behavior (Reference KnightKnight 1992; Reference WahlbeckWahlbeck 1998). Here, insofar as individuals comply with the legal rule, there should be less litigation in the first place (Reference WahlbeckWahlbeck 1998). Beyond this initial impact, after the Supreme Court has announced a new legal rule, lower court judges are more likely to support the Court's rule than to support a revision to the rule (Reference WahlbeckWahlbeck 1998), at least in some issue areas. This reduces the uncertainty of litigants, as lower court judges comply with the rule laid out by Supreme Court decisions (Reference MerrymanMerryman 1954). When uncertainty over outcomes is reduced, as in after a Supreme Court decision, litigants garner more accurate assessments of their probability of success through litigation (Reference MerrymanMerryman 1954). With a clear understanding of their probability of success, litigants unlikely to succeed are bound to refrain from filing or appealing their cases (Reference Priest and KleinPriest & Klein 1984). Therefore, to the extent that Supreme Court decisions clarify the law, they should also lead to less litigation.
Taken together, this perspective suggests that justices seek primarily to quell rising tides of litigation in areas where there is confusion or dissensus, areas where decisions will have the most general impact. By acting, the Court establishes legal rules that lower uncertainty throughout the legal universe, as lower courts comply with the rule, and gain a clearer guidelines for instances of non-compliance. This compliance and clarity reduces legal uncertainty and litigation, and therefore reduces the total number of cases within a policy area. In sum, this perspective predicts that when the Supreme Court addresses an issue, there will be less subsequent attention to that issue in lower federal courts.
The Interest Group Perspective and Mobilizing Litigants
Quite separate from this conventional perspective, an interest group perspective on responses to signals from the Supreme Court has recently emerged. This theory has its roots in foundational works on the behavior of interest groups across institutional settings (Reference TrumanTruman 1951) including the Supreme Court (Reference ShapiroShapiro 1964). Indeed, interest groups are active participants in the activities of the judicial branch, with amicus briefs filed in over 90% of Supreme Court cases decided in recent years (Reference CollinsCollins 2007). Scholars studying the court have focused on a variety of interest group influences including decisions granting certiorari, final judgments and influences on the composition of the courts (see Reference Caldeira, Hojnacki and WrightCaldeira, Hojnacki, & Wright 2000; Reference Epstein and RowlandEpstein & Rowland 1991; Reference KobylkaKobylka 1987; Reference McGuire and CaldeiraMcGuire & Caldeira 1993). Of particular note, research has indicated that the way groups present legal rules to courts can influence judicial decisionmaking (Reference Epstein and KobylkaEpstein & Kobylka 1992). This is in line with research noting that the policy-making power of courts relies on litigants and groups being willing to bring cases (Reference EppEpp 1998).
Recently, scholars have proposed that when the Supreme Court acts, its decisions impact group litigation patterns. Whereas the legal perspective suggests reduced incentives to litigate after the Supreme Court has addressed an issue, the interest group perspective suggests that the unique incentives of interest groups lead to additional and better-framed litigation. Because some litigants, and particularly groups, are policy-minded, these litigants pay attention to the signals in Supreme Court decisions about how arguments could be framed in future cases so as to garner their preferred policy result (Reference Baird and JacobiBaird & Jacobi 2009b). From this perspective, the sitting justices impact the future agenda because “(t)he incentive to support litigation in particular policy areas varies over time in accordance with litigants’ changing perception of Supreme Court justices’ policy priorities” (Reference BairdBaird 2007: 4). Increases in attention to an issue signals the Court's desire to address that issue, and subsequently leads to temporary increases in the Court's attention within that area as litigants bring additional better-framed cases. The result is “… that some time after Justices signal their interest in hearing particular types of cases, a significantly higher number of those cases will be brought to the courts and result in case outcomes” (Reference Baird and JacobiBaird & Jacobi 2009b: 222–23).
Previous research has provided corroborative evidence for the Supreme Court's ability to manipulate this reaction. Reference BairdBaird (2004, Reference Baird2007) supported the notion of litigant mobilization through evidence of increased Supreme Court attention to policy areas anywhere from 4 to 6 years, and in a sample of published courts of appeals opinions four years, after the Court had indicated that policy area as a priority. This lagged effect is presumed to be consistent with the period of time necessary for cases and issues to “percolate” through the lower federal courts to the Supreme Court. In yet other research, the Supreme Court paid additional attention to federalism in years after dissents argued that cases should be decided on federal-state power relations (Reference Baird and JacobiBaird & Jacobi 2009a,Reference Baird and Jacobib).
While this signaling model has been the subject of much scholarly attention, whether litigants and interest groups are driving this process is still in dispute. There is contradictory evidence as to whether litigants change their behavior in filing amicus curiae briefs within an issue area after Supreme Court decisions in that issue area (Reference BairdBaird 2007; Reference PetersPeters 2007).
In addition, Reference PetersPeters (2007) shows that there is no increased attention after Supreme Court activity in other venues, such as law reviews, where the new frames would be expected to be discussed. Finally, the model does not account for the counter-mobilization of litigants opposing judicial attention (Reference Solowiej and CollinsSolowiej & Collins 2009).
Despite these disputes, the emerging interest group perspective provides a set of clear empirical expectations about the effects of a Supreme Court ruling on the lower federal courts. Specifically, it suggests—in contrast to the conventional wisdom—that after the Supreme Court pays particular attention to an issue area, the level of attention paid by lower courts to that issue area will temporarily increase.
Can Both Operate?
But while the predictions of the two perspectives diverge, the dynamics underling those predictions are not necessarily at odds. It is conceivable for Supreme Court activity to settle issues of law, and therefore lower the overall amount of issue attention in lower federal courts, while at the same time providing signals to litigants about a particular re-framing of an argument to which the Court would be receptive, or some secondary considerations, which have yet to be addressed. The result would be fewer cases in the aggregate within the issue area in question while, simultaneously, the remaining cases could be the subject of interest to members of the litigant community, particularly if they present new arguments in line with the Court's signals. Consistent with the conventional perspective, the law—in general—would be settled, with a concomitant decrease in overall issue attention to the area the Supreme Court has addressed. However, consistent with the interest group perspective, there would also be additional litigant mobilization on the issue, with groups bringing or supporting cases that are particularly apt policy vehicles.
Reference KleinKlein (2002) offers evidence that such a dynamic may exist. In his study of the U.S. courts of appeals, he notes that, in some issue areas, judges were forced to create legal rules in order to interpret recent Supreme Court rulings (Reference KleinKlein 2002: 59). So while the Supreme Court's attention to an issue area may have settled some aspects of the law, interpreting and implementing the decision may nonetheless require new legal rules. Sophisticated litigants and interest groups are, for judges, valuable participants in the development of these new legal rules (Reference Epstein and KobylkaEpstein & Kobylka 1992).
Overall, in this view the Court's decisions resolve areas of unsettled law, thereby decreasing the number of cases in an issue area in the lower courts, while also signaling additional concerns which, when litigated with the help of sophisticated interest group litigants, lead to yet further refinement of the legal rule. The remaining cases thus have additional interest group participation, and are particularly complex, as they institute a new legal rule. This view thus predicts fewer cases within an issue area after the Supreme Court acts on that issue, but also increasingly complex cases after the Supreme Court acts on that issue.
Data and Measurement
The unit of analysis in judicial agenda studies are court issue areas, with the measurement of the issue attention of courts typically accomplished through counts of cases, classified by issue area (Reference Baumgartner, Gold, Baumgartner and JonesBaumgartner & Gold 2002; Reference PacellePacelle 1991, Reference Pacelle and Epstein1995). Scholars seeking to understand the agenda of the lower federal courts (e.g., Reference HurwitzHurwitz 2006) have, in general, generated these case counts by aggregating cases in the United States Courts of Appeals Database, an extensive data collection that contains information on a sample of courts of appeals cases with full opinions published in the United States Reports.
But changes in the number of published opinions will not fully differentiate each of the perspectives. Opinions are published when they have precedential value, with that decision left largely to the discretion of the court. Counts of published opinions, therefore, do not reflect litigant mobilization as much as they reflect changes in the level of precedential change on an issue, and as such represent only part of what is required to differentiate the conventional and interest group perspectives. Moreover, the decision on publication is left to lower court judges, therefore published opinions offer a signal of how lower court judges prioritize issue areas, rather than litigants. Given what has been mentioned earlier, I employ three dependent variables, which reflect different aspects of agenda attention in the federal courts. First, I need to capture the overall rate of litigation, including cases disposed at all stages in the legal process, and which are predicted to decrease after Supreme Court decisions. Second, I need to measure the amount of precedential change, reflected in the decisions of lower court judges to publish opinions in specific cases. Finally, I need to capture interest group participation in this process, crucial as it is to the interest group perspective.
To capture these aspects of judicial attention, I employ three different measures of issue attention in the lower federal courts, each of which provides slightly different evidence of the extent to which the law is settled or being developed in the lower federal courts. These three measures are counts of all cases by issue area, a count of all published opinions by issue area, and counts of published opinions with amicus curiae participation.
In order to obtain data on the total amount of litigation in the federal courts, I utilized filings on case terminations made available from the AO of the United States Courts (Federal Judicial Center. Conducted by the Federal Judicial Center n.d.).Footnote 2 These data include information on all cases within the federal courts; as such, they reflect the priorities accorded to issue areas for all categories of litigants. Moreover, the data include cases of all types, including those with both published and unpublished opinions, making this the first agenda study to have a comprehensive view of the judicial agenda.Footnote 3
By examining all federal court cases, rather than solely those with published opinions, I gain insight into the influence of important Supreme Court decisions on the overall lower court agenda and the dynamics of issue attention across all federal court activity.
What do the perspectives mentioned earlier suggest happens after the Supreme Court acts to patterns of litigation in the federal courts? The conventional, legal view predicts there will be a decrease in the amount of litigation in an issue area after the Supreme Court has acted in that issue area. Contrary to this prediction, the interest group perspective suggests that temporary increases in litigation will follow Supreme Court attention as signaling results in, “… more cases being brought in a policy area when Justices signal that the policy area is a judicial priority” (Reference Baird and JacobiBaird & Jacobi 2009b: 217).
The second measure is a count of the published courts of appeals opinions only, also by issue area. Courts of appeals opinions, as Reference SongerSonger (1990) reviews, are published by different standards across circuit courts, issue areas, judges, and litigants. On average, cases are not disposed with published opinions; following the recommendations of Judicial Conference of the United States in 1964, only cases with precedential value should be published (Reference Davis and SongerDavis & Songer 1989). Therefore, these cases indicate new approaches, applications, or developments of legal rules. To the extent that the presiding judge(s) believed a new frame was being presented, they would deem publication of an opinion necessary to address the issue.
The conventional perspective predicts a decrease in the number of published opinions in an issue area after the Supreme Court has acted in that issue area. To the extent Supreme Court opinions establish a legal rule intended to guide future behavior, as the conventional legal perspective holds, lower court judges should respond by publishing fewer opinions on the issue. Proponents of the legal perspective suggest this pattern emerges as less litigation arises on the issue in the first place, and because judges find less precedential value in cases as they apply the Supreme Court's rule (Reference WahlbeckWahlbeck 1998).
In contrast, the interest group perspective predicts an increase in the number of published opinions, and indeed previous research for this perspective has documented such an increase opinions in the courts of appeals as corroborative evidence for the interest group perspective (Reference BairdBaird 2004, Reference Baird2007). According to this view, the additional published opinions arise as policy-minded litigants identify appropriate vehicles to limit or expand on the Supreme Court's precedent (Reference BairdBaird 2004, Reference Baird2007). Their attention provides indications of the need for attention to lower court judges, who may then seek to further clarify points unclear in the Supreme Court precedent (Reference KleinKlein 2002). Therefore, this perspective holds that, after Supreme Court attention to an issue, lower courts will publish more opinions on that same issue.
Finally, the interest group perspective hinges on the mobilization of policy-minded litigants, in particular interest groups, in the federal courts after Supreme Court attention to issues. In recent years, there have been marked improvements in our understanding of amicus brief filings in the federal courts, particular in regard to factors motivating groups to participate as amici (Reference Collins and MartinekCollins & Martinek 2011; Reference Epstein, Knight, Clayton and GillmanEpstein & Knight 1999; Reference HansfordHansford 2004a,Reference Hansfordb; Reference KobylkaKobylka 1987; Reference MartinekMartinek 2006; Reference Solowiej and CollinsSolowiej & Collins 2009). From this research, there is strong evidence that the intention of groups in filing amicus briefs is primarily to shape public policy.
Here, the perspectives again offer divergent expectations. Within the judicial hierarchy, the conventional view suggests interest groups will be less active in the courts post-increases in Supreme Court attention, as the Court settles areas of unsettled law. The conventional perspective, on the contrary, suggests that sophisticated legal counsel of interest groups can respond to specific signals in Supreme Court opinions and seek out the appropriate case vehicles on which they mobilize (Reference BairdBaird 2004, Reference Baird2007; Reference Baird and JacobiBaird & Jacobi 2009a,Reference Baird and Jacobib). Interest group participation in the federal courts, given their specific concern in shaping public policy and their hypothesized response to Supreme Court attention, is therefore integral to testing the influence of Supreme Court decisions on issue attention in lower federal courts. While interest groups have a variety of potential avenues for influence, filing an amicus curiae brief is an increasingly common choice (Reference CollinsCollins 2008; Reference Collins and MartinekCollins & Martinek 2011).Footnote 4
In order to understand how issue attention changes across each of these three measures, I need to classify each federal court case into issue categories. This same task has been carried out across a number of different policy-making institutions, including the Supreme Court, by the Policy Agendas (PA) Project,Footnote 5 a large-scale effort to classify institutional or government actor attention into 19 major topic (or issue) codes. In categorizing the lower court measures, I adopt these 19 major issue categories.Footnote 6 The categories are assigned based on the policy content of the measure. For example, the landmark case of Lawrence v. Texas,Footnote 7 where the Supreme Court ruled a Texas law banning sodomy violated the Due Process Clause, is coded as “Civil Rights, Minority Issues, and Civil Liberties.” I choose to employ the PA scheme for a number of reasons. First, doing so allows for the inclusion of other institutions as predictors of issue attention. Institutional agendas are inextricably linked in American government, and analysis of agenda dynamics have uncovered evidence for, and have encouraged subsequent research to consider, cross-institutional influences in agenda studies (Reference Flemming, Bohte and WoodFlemming, Bohte, & Wood 1997). Beyond the influence of other institutions, the classification scheme employed by the principal alternative—the Supreme Court Database (SCDB)Footnote 8 —is focused on the disposition of individual cases, and thus includes as issue codes particularly procedural concerns (i.e., mootness) with little relation to attention dynamics, the subject of study here. Finally, the issue codes of the SCDB have themselves come under criticism as being inaccurate (Reference Harvey and WoodruffHarvey & Woodruff 2011; Reference ShapiroShapiro 2009). Given these problems, I utilize the PAs scheme.
To create measures of lower court issue attention, I use the text of Supreme Court opinions, acquired from Public.Resource.Org, and the PA codes in a supervised machine learning algorithm. Supervised learning methods utilize a subset of hand-coded training data to train a model which is then evaluated on a subset of hand-coded test data. The goal is to select a model that best predicts the test data. Here, I used as training and evaluation data the PA's hand-coded Supreme Court opinions. I trained naive Bayes, maximum entropy, and decision tree classifiers, which I then used in a simple ensemble classification algorithm (Reference Hastie, Tibshirani and FriedmanHastie, Tibshirani, & Friedman 2008) to predict the issue areas of the published courts of appeals opinions.
I used the classifications of published opinions in lower federal courts, in concert with the AO of the Federal Courts data on all federal court cases, to classify all district and courts of appeals cases. To determine the issue content of these cases, I matched cases from the AO with their published opinions, and therefore to the issue assigned to the case through the ensemble classification algorithm. I then used the nature of suit (civil) or offense (criminal) codes, which the AO assigns to every case, to map the PA issue codes to the issue codes utilized by the AO. The result is a mapping of the AO nature of suit and offense codes to the PA issue coding scheme that provides a consistent.
With the issue classifications in hand, I created three distinct measures of issue attention within the federal courts. The first, the overall agenda, is an aggregate of all cases within an issue area in a federal district court or court of appeals, during a year.Footnote 9 The second, the published agenda, is an aggregate of all published opinions within an issue area for a given court, during a year. Finally, I measure interest group participation as the number of published opinions with at least one amicus curiae, within an issue area for a given court, during a year.Footnote 10 While ideally, interest group participation would be measured as the number of cases actually sponsored by interest groups, that information is unavailable on any large-scale basis. Instead, I employ the presence of amicus curiae as a proxy, as they offer an indication of interest group participation and interest group efforts to influence public policy.
Supreme Court Attention
In each of the perspectives outlined earlier, the key influence on lower court issue attention is the attention given to that issue by the Supreme Court. Two options exist for this measurement. The first, and obvious, option is simply the total count of Supreme Court cases within an issue area during a year. However, this measure could potentially lead to masking the signals which litigants receive, and therefore biasing the analysis toward the legal perspective.
As an alternative, previous research has utilized counts of salient, or important, court cases within policy areas, in order to capture signals of areas the Court considers a priority (e.g., Reference BairdBaird 2004; Reference PetersPeters 2007). If litigants respond to what they believe are the priorities of the Supreme Court justices, as the interest group perspective suggests, then the measure should capture the litigant's perceptions of the Supreme Court's priorities. Using the measure of salient cases by policy area, as suggested by previous research on the interest group perspective, satisfies this requirement for the interest group perspective. Further, for the conventional legal perspective, important cases should still precede fewer cases, although important cases are likely to understate the influence of the Supreme Court.
Given the strengths and weaknesses of each measure, I estimate separate models with the different operationalizations of Supreme Court attention. The measure of overall Supreme Court cases is the number of all court cases within an issue area during a year. The measure of salient Supreme Court cases is the number of salient court cases within an issue area during a year. I identify the subset of important or salient cases Reference Epstein and SegalEpstein and Segal (2000) measure of case salience; this measure is simply whether or not a case appears on the front page of the New York Times the day after it is decided, and is the featured case in that article. The measure thus classifies a case like City of Boerne v. Flores Footnote 11 as salient, while not classifying a case like Edwards v. Balisok. Footnote 12 Advantageously, the New York Times measure has been utilized in similar research previously (Reference BairdBaird 2004; Reference PetersPeters 2007), and primarily captures contemporaneous political salience, which is the most likely to be perceived by litigants and interest groups.
Legislation
While Supreme Court attention is almost certainly an important driver of lower court issue attention, it is also not the only influential factor. Previous research suggests that any model of judicial attention must account for congressional issue attention, another important cross-institutional influence on the courts (Reference Baird and HurwitzBaird & Hurwitz 2006; Epstein, Segal, & Reference Epstein, Segal and VictorVictor 2002; Flemming, Wood, & Reference Flemming, Wood and BohteBohte 1999). While the directionality of this relationship is unclear (Reference Flemming, Wood and BohteFlemming, Wood, & Bohte 1999), there is ample reason to suspect that congressional activity will influence issue attention in the lower federal courts. In fact, much of what the courts do lies in interpreting Congress-made law (Reference Johnson and CanonJohnson & Canon 1984). Fundamentally, Congress, through the passage of laws, determines the content of cases before the courts (Reference Flemming, Wood and BohteFlemming, Wood, & Bohte 1999). As an example, Reference Baumgartner and JonesBaumgartner and Jones (1991) point to the National Environmental Protection Act of 1969, which expanded access to the courts and “forced [courts] to give greater consideration to certain aspects of environmental policy which had been ignored in the past” (1049). Therefore, as legislation is passed, court cases arise dealing with scope, issues of interpretation and implementation, or a multitude of other issues (Reference ZemansZemans 1983). As Reference PacellePacelle (1991) notes, the litigation response ultimately could even drive changes in the Supreme Court's agenda.
Thus, I need to capture the influence of Congressional attention, particularly legislation, on the judicial agenda. To measure the impact of legislation, I use an aggregated count of the number of public laws passed within each of the 19 major PA issue categories during each year. Higher values thus indicate that Congress was more active within an issue area during a particular year, while lower values indicate Congress was less active on that issue area during a year.
Executive
Beyond Congress, any model of agenda-setting must also account for the executive branch, as “no other single actor can focus attention as clearly, or change the motivations of such a great number of other actors, as the president” (Reference Baumgartner and JonesBaumgartner & Jones 1993: 241). While empirical support for the influence of the president on the agenda is mixed (Reference Edwards and WoodEdwards & Wood 1999; Flemming, Wood, & Reference Flemming, Wood and BohteBohte 1999), they have a clear and direct potential influence on the lower federal court agenda through the Department of Justice and the office of the solicitor general. The president can directly influence the subset of the case, which comes before the judiciary, as the office of the solicitor general chooses to prioritize certain issue areas. Moreover, in recent terms, the office of the solicitor general has been employed explicitly sought as an avenue for presidential issue attention influence through the filing of “agenda” cases (Reference PacellePacelle 2003; Reference SalokarSalokar 1992; Reference WohlfarthWohlfarth 2009). Finally, research on criminal justice policy suggests that the president's rhetoric actually influences the implementation process (Reference MeierMeier 1994; Reference Yates and WhitfordYates & Whitford 2009). On some issues, then, presidential attention is likely to predict subsequent lower court attention.
The attention and priorities of the president are notoriously difficult to operationalize (Reference Edwards and WoodEdwards & Wood 1999). Here, my concern is in capturing the president's publicly expressed priorities, as those are the priorities most likely to be reflected in broad-scale influence on the judicial agenda. Therefore, to measure executive attention, I aggregate the PA data on the number of mentions a particular issue area receives in the President's annual state of the union address. The measure thus captures the policy priorities of the president, as the president is willing to publicly express them.
Model Specification
The structure of the data is such that each instance of the outcome of interest is a measure within a policy area, within a court, during a year.Footnote 13 To account for those multiple groupings, I estimate a multilevel model with mixed effects for court and policy area, which allows coefficient estimates to vary by unit while also avoid the large and unreasonable variances across estimates of a “no-pooling approach” (Reference Gelman and HillGelman & Hill 2007). Between the extremes of large variance or no variance across court-issue areas, multi-level modeling will converge toward the correct pooling specification. It therefore offers an optimal approach to the structure of this data and the subject of interest.
Variance across issue types in Supreme Court activity has been documented and controlled for in previous studies (Reference HurwitzHurwitz 2006; Reference LinkLink 1995). I also expect there to be variation in caseloads within issue areas by court. The differences within these districts and circuits, the differences in issue foci and the differences in the involved attorneys would all lend credence to an expectation of variance across issue and court. Therefore, in the multilevel specification I estimate random intercepts for both policy area and court (whether district courts or courts of appeals).Footnote 14
As in similar previous research (Reference BairdBaird 2004, Reference Baird2007; Reference PetersPeters 2007), I estimate an autoregressive distributed lag model of the effect of Supreme Court attention. I include five lags of each independent variable, both to be consistent with this same prior research, and because their findings have suggested influence from Supreme Court attention across this entire period. Additionally, a lagged dependent variable is included to account for serial correlation, a precaution still necessary in multi-level modeling approaches (Reference ShorShor et al. 2007). The multilevel model is thus:
where j indexes the issue and k the district or circuit. The respective measure of lower court judicial attention within a court issue area is given by yjkt.Footnote 15 The varying intercepts are given by αjk, and φo is the coefficient on the lagged dependent variable. The expression β 1:5 represents the coefficient estimate on the lags of the number of important Supreme Court cases, while β 6:10 represent the coefficient estimates on lags of the number of public laws passed and β 11:15 represent the same for the number of State of the Union mentions. Varying intercepts are estimated for court and policy area to account for court- and issue-specific effects.
Model of Litigant Attention
I begin with the results of the models of all litigation in the lower federal courts, or the measure of litigant mobilization. How do litigants, motivated by the outcome of the case at hand, respond to Supreme Court attention? Table 1 contains the results of models of all litigation in the federal district courts and courts of appeals. Across the five lags of both measures of Supreme Court attention, we see a generally negative relationship with subsequent litigation in federal district courts.
Table 1. Models of All Litigation in the Federal District Courts (1970–2005) and Courts of Appeals (1971–2006)
Note: *indicates p < 0.05 (two-tailed). For both models, N = 5,168, groups: policy = 19, circuits = 12.
Figure 1 makes the dynamics clearer. In Figure 1, I plot the coefficient estimates for each of the five lags of the respective measure of Supreme Court attention. In the federal district courts, there is stark evidence for the legal perspective in the response of all litigants. Increased Supreme Court attention to an issue area, whether measured as the total number of cases the Court decides or as only the salient cases, correlates with less district court attention to that issue in nearly all subsequent years under analysis. Moreover, increases in the overall number of Supreme Court cases in an issue area correlates with less attention in the courts of appeals to that issue 1 year after the increase in Supreme Court attention.
Figure 1. Fixed Effects Estimates for Influence of Supreme Court Issue Attention on Overall Federal Court Caseloads.
The long-run cumulative effect in distributed lag models with a lagged dependent variable can be calculated as 
. For example, the long-run cumulative effect of Supreme Court attention at one lag, measured as salient cases, on the issue attention of district courts equals −0.115. Indeed, at each lag, the long-run cumulative effect is greater than the reported immediate-effect coefficients. In all, litigants, in general, pay less attention to issue areas which the Supreme Court addresses.
Thus, it appears from this particular analysis that the Supreme Court does exactly what justices claim; the Court addresses issues with the goal of settling areas of the law where uncertainty is high. Litigants, motivated by the likelihood of winning their individual cases, have less incentive to litigate as Supreme Court decisions clarify or establish legal rules which help to guide behavior. Over time, Supreme Court decisions lead to less litigation. In sum, the Court fulfills its role as the final arbiter in the legal system, resolving difficult questions. This dynamic suggests support for the conventional, legal perspective.
Model of Published Opinions
While Supreme Court attention stems overall levels of litigation, how does it shape the priorities of judges and interest groups? I now move to an analysis of published opinions, which reflect changes in the attention of judges in the lower federal courts. The results of the models of published opinions in the lower federal courts are presented in Table 2. Here, the results present a different picture. Through published opinions, judges address an issue area more often after the Supreme Court addresses that issue area, which is consistent with the interest group perspective.
Table 2. Models of Published Opinions in the Federal District Courts (1950–2000) and Courts of Appeals (1950–1996)
Note: *indicates p < 0.05 (two-tailed). For both models, N = 8,759, groups: policy = 19, circuits = 12.
In contrast with the attention of all litigants, the attention of judges actually increases 3–5 years after increases in Supreme Court attention. In models with all Supreme Court cases as the operationalization of Supreme Court attention, the dynamics are even suggestive of the process of judicial response, with the relationship first statistically significant in the district courts 3 years after an increase in Supreme Court attention, and then 4 and 5 years after Supreme Court attention in the courts of appeals. Again, I present the results as ropeladder plots of the coefficient estimates in Figure 2. The dynamics of the relationship are clear; there is little change in the attention of lower court judges to particular policy areas until 3 years after the decision, at which point judges pay additional attention to the issue area. This additional attention then tapers off, however. These observed dynamics are entirely consistent with prior research (Reference BairdBaird 2004, Reference Baird2007) in support of the interest group perspective.
Figure 2. Fixed Effects Estimates for Influence of Supreme Court Issue Attention on Cases with Published Opinions in Lower Federal Courts.
Judges, attentive to the need for modifications at the edges of Supreme Court decisions (Reference KleinKlein 2002), exert additional influence in issue areas, which the Court addresses. For judges supportive of the legal rule established in the Supreme Court opinion, this could mean they seek to expand the rule through their published opinions. For judges in disagreement, they can seek to limit the applicability of the rule. In either case, the Supreme Court's attention to an issue encourages additional attention from judges in lower courts. The Court is thus able to engage in a dialog with lower federal courts over the state of legal policy, encouraging additional issue percolation in lower federal courts (Reference Clark and KastellecClark & Kastellec 2012) and thus opening the door to greater judicial influence on public policy.
Model of Interest Group Attention
Across these first two actors, then, we have evidence that suggests that Supreme Court attention to an issue area both discourages additional litigation in that issue area, while also encouraging additional attention within that issue area among judges. An important repeat player in the federal courts, interest groups, is our final actor. Recall that interest groups are particularly concerned with public policy, and are hypothesized to be at the center of the litigant signal model. As repeat players, with goals most closely aligning with those of judges, the results mentioned earlier suggest that interest groups will respond positively to Supreme Court attention.
The results are presented in Table 3. Here, we see results that are generally consistent with the interest group perspective. Increased Supreme Court attention to an issue is correlated with additional attention to the same issue in lower federal courts between 3 and 5 years later in both federal district courts and courts of appeals, consistent with the expectations of the interest group perspective. As mentioned earlier, the coefficient estimates on Supreme Court attention are plotted in Figure 3 for ease of interpretability. Interest groups are more involved in cases before the lower federal courts in issue areas which the Supreme Court has addressed 3–5 years earlier.
Figure 3. Fixed Effects Estimates for Influence of Supreme Court Issue Attention on Cases with at Least One Amicus Curiae.
Table 3. Models of Cases with Amicus Curiae in the Federal District Courts (1950–2000) and Courts of Appeals (1950–1996)
Note: *indicates p < 0.05 (two-tailed). For both models, N = 8,759, groups: policy = 19, circuits = 12.
However, we also see for the first time a differentiation in results across the models with different measures of Supreme Court attention. Prior to the relationships described earlier, an increase in the overall Supreme Court cases within an issue area leads to immediate (1 or 2 years later) decreases in the number of cases with an amicus curiae in the lower federal courts. This decrease, consistent with the legal perspective, suggests that interest groups are not as active on an issue in the federal courts immediately after Supreme Court attention to that issue. The difference is likely attributable to the heightened impact of salient decisions, which offsets any settling effect from Supreme Court opinions.
Across these analyses, the evidence suggests dynamics consistent with both legal and interest group perspectives. Notable institutional actors—judges and interest groups—in the federal courts pay additional attention to issues 3–5 years after the Supreme Court addresses them. However, if we consider all litigants or the immediate responses of interest groups, then the pattern is actually the opposite, with less attention devoted to issues overall in lower federal courts after the Supreme Court addresses that issue.
Summary and Conclusion
What happens in the federal courts after the Supreme Court acts? The conventional wisdom suggests that the purpose of Supreme Court decisions is to settle areas of the law. Therefore, after the Supreme Court acts in an issue area, there should be less litigation within the federal courts in that issue area. On the other hand, an interest group perspective suggests that when the Supreme Court acts, they also send signals to litigants about potential avenues for future litigation. As litigants respond to those signals, they temporarily increase the amount of attention to that issue area in the federal courts. At the confluence of these two perspectives are divergent expectations for what happens in the federal courts after the Supreme Court acts.
In order to resolve this discrepancy, I utilized information on all case filings in the courts of appeals and district courts over an extended period of time, and all published opinions from the courts of appeals over an even longer span of time. By adopting machine learning methods for supervised classification, I created three separate measures of issue attention in the lower federal courts: total caseloads by issue area, published opinions by issue area, and amicus curiae activity by issue area. These measures are directly comparable across all levels of the judiciary, and with measures of congressional and presidential issue attention. In totality, this provided the most comprehensive picture to date of issue attention in the federal courts.
Consistent with the conventional legal perspective, I found that overall issue attention, measured by levels of litigation within issue areas, decreases after Supreme Court activity increases in the issue area. In other words, Supreme Court decisions, in general, work to settle the law within an issue area. Moreover, 1–2 years after increases in the Supreme Court's overall attention to an issue, interest groups are less active on that issue in lower federal courts. Together, these decreases provided evidence opposing a purely interest group perspective that the Supreme Court signals litigants to mobilize. Contrary to these decreases, the number of published opinions increased, and, 3–5 years later, the activity of interest groups increased. These increases therefore provided evidence against a purely legal perspective. Opinions are only published if they have precedential value, implying that the opinions address a new legal argument or circumstance. Thus, neither perspective prevalent in current research provides a complete answer to what happens when the Supreme Court acts. Instead, the answer is a synthesis of these two perspectives: while, in general, the Supreme Court's attention to an issue decreases attention to that issue in the federal courts, the Court's attention also opens the door to new legal arguments which are subsequently debated in the few remaining cases in the issue area.
In demonstrating this connection between the attention of the Supreme Court and subsequent attention in lower federal courts, these findings have implications for agenda setting, policy making, and legal mobilization. Agenda setting is well-established as a crucial step in the formation of public policy (e.g., Reference Bachrach and BaratzBachrach & Baratz 1962), and the research reported here demonstrates the Supreme Court's influence over the scope of policy emanating from the judiciary. The Court resolves difficult questions with widespread impact on the legal system, often establishing legal rules which preclude litigation within the issue area in which the Court has acted. While the justices on the Supreme Court can accomplish their stated goal of stemming tides of litigation, the Court also engages in a policy-making dialog with the lower federal courts. Policy-minded justices are therefore able to fulfill both their perceived legal duty while also pursuing their policy preferences. In all, by discouraging litigant mobilization, but encouraging attention from judges—in the form of published opinions—and interest groups—in the form of amicus curiae participation—the Court exerts considerable and previously understudied influence on the judiciary.
In this article, I have documented the important influence of the Supreme Court's attention on issue attention in lower federal courts. An important avenue for future research to explore is variation in this influence across qualitative differences in the Supreme Court's attention. For instance, it is possible that the clarity of Supreme Court opinions influences the amount of subsequent attention that lower court judges must devote to a particular policy. Reference KleinKlein's (2002) research on the courts of appeals certainly suggests that such differences in the attention may exist.
I leave this question to future research. For now, the goal of this article was simply to determine whether the impact of Supreme Court decisions was to decrease attention to issues, as the legal perspective suggests, or to increase attention to issues, as the interest group perspective suggests. The answer, it seems, lies somewhere in between. On the whole, additional Supreme Court attention leads to fewer cases within an issue area. Of those remaining cases, however, more have published opinions, and those published opinions are longer, indicating that new legal arguments are being parsed in the lower federal courts.
