1.1 The Unusual Influence of Donald Trump on the Courts

The presidency of Donald Trump will be remembered for many reasons, but perhaps its most enduring legacy will be its impact on the federal judiciary. Trump appointed more Supreme Court justices than any president since Ronald Reagan, and the most in a single term since Richard Nixon. His influence was not limited to the Supreme Court. In just four years, he appointed 27% of active district court judges, and 30% of active courts of appeals judges. His appointees are significantly more conservative and less diverse than those of both his Republican and Democratic predecessors.Footnote ¹⁰ They are also young – at 47, the average age of Trump’s nominees was lower than that of any president in at least a century.Footnote ¹¹ Their youth ensures that many of them will serve for decades. These appointments have the potential to dramatically reshape the types of cases that come before the courts, the legal doctrine the courts create, and which litigants view the courts as a viable policymaking venue for many years to come.

Who holds both elected and appointed public office is consequential. Research on the composition of government shows that who is in office affects everything from citizens’ trust in our institutions (see, e.g., Reference GayGay, 2002) to the issues that make it on to those institutions’ agendas (see, e.g., Reference Reingold, Haynie and WidnerReingold, Haynie, & Widner, 2020) to the substantive policy decisions made by those institutions (see, e.g., Reference BoydBoyd, 2016). Less is known, however, about whether and how characteristics of those who hold office in a particular institution affect who participates in advocacy before that institution.

Representation and participation are particularly important with respect to the courts. While all branches of the US government are responsive to public pressures to some degree, officials in the legislative and executive branches are able to initiate policy activities based on their own interests, life experiences, or policy preferences (Reference BurdenBurden, 2007). Federal courts, on the other hand, are limited by the “case or controversy” requirement of Article III of the Constitution. This requirement limits courts’ decision-making authority to situations in which someone has suffered or will imminently suffer an injury or legal impairment, and courts can only weigh in once that party properly files before the correct court. This makes the judicial branch a reactionary institution, rather than a proactive one (see, e.g., Reference BandesBandes, 1990). Because courts’ policymaking opportunities are dependent on litigants to present issues for them to resolve, litigants play an essential role in agenda-setting for the judicial branch. However, this agenda-setting influence flows in both directions. Although courts can only take case litigants bring to them, judges can signal to potential litigants their interest in taking up a particular type of case (Reference BairdBaird, 2004). Further, the very presence on the bench of judges whose policy preferences are known or presumed can lead litigants to expect the Court to be a more or less receptive audience for their claims.Footnote ¹²

While a complete accounting of Trump’s influence on the judiciary would consider judicial confirmations at all levels of the federal judicial hierarchy, we focus our attention here on Supreme Court confirmations during Trump’s presidency. Supreme Court confirmations have become highly partisan, contentious affairs (Reference Armaly and LaneArmaly & Lane, 2023; Reference Cameron, Kastellec and ParkCameron, Kastellec, & Park, 2013; Reference Collins Jr, Ringhand and BoydCollins, Ringhand, & Boyd, 2023). They capture the attention of the public and the media. Moreover, particularly during the Trump presidency, controversial confirmations substantially eroded public support for the Court and the nominees specifically (Reference Carrington and FrenchCarrington & French, 2021; Reference Rogowski and StoneRogowski & Stone, 2021). Because Supreme Court confirmations are salient political affairs that are well-covered by the media, interested parties in litigation are likely to pay attention to confirmations as signals of the Supreme Court’s receptiveness to particular policy areas or legal strategies. The interaction of these strategies and the justices’ preferences have significant implications for national policymaking.

1.2 Judicial Turnover, Case Selection, and Outcomes

Judicial turnover can be impactful at all levels of the federal judiciary, but it has the most significant effect at the Supreme Court. While lower courts have mandatory jurisdiction – meaning that they must respond to all properly filed cases – the US Supreme Court, for the most part,Footnote ¹³ chooses the cases it wishes to hear. The Court is very selective; each year it receives between 5,000 and 7,000 requests for review and grants only about 80 of these a full hearing (Reference Lane and BlackE. Lane & Black, 2017).Footnote ¹⁴ After reviewing petitions for certiorari, which are parties’ official requests for the Court to hear a case, members of the Court meet and vote on whether to grant review. It takes four members of the Court to grant review.Footnote ¹⁵ As new justices come onto the Court, they inevitably affect this process. Justices vote independently, bringing their own preferences and beliefs about what cases are important, but they also vote strategically, considering how their colleagues might vote at the merits stage (Reference Black and OwensBlack & Owens, 2009; Reference PerryPerry, 1991). Thus, a new justice affects the Court’s agenda both through their own votes and through their impact on the strategic calculus of other justices.

Figure 1, which is drawn from the Supreme Court Database (Reference Spaeth, Epstein and MartinSpaeth et al., 2021), shows the number of cases that the Supreme Court has reviewed in key issue areas over time. The most obvious trend is the general decline in the number of cases the Court has taken. In the 1980 term, the Supreme Court heard 150 cases, but starting in the mid-1980s, its caseload declined sharply. By 2000, the number of cases the Court took in a term had fallen to 86. Since then, the number has consistently been even lower, and has often been under 70 cases a term.Footnote ¹⁶ A closer look at the figure also reveals some interesting changes in the types of cases the Court takes. While all issue areas decline and all have occasional spikes over time, issue areas like criminal procedure and federalism show a steep and fairly consistent decline, whereas economic activity cases decline with other case types until the late 1990s and then rebound and level off at a higher proportion of the docket in the 2000s. The rebound happens right around the time that Chief Justice John Roberts – a justice who worked for Ronald Reagan and championed economic deregulation – took the bench. In the past decade, the number of economic activity cases per year has frequently surpassed the number of criminal procedure cases, the issue area that had traditionally dominated the Court’s docket.

Figure 1 Issue area of SCOTUS cases granted cert over time, 1980 to 2021

Table 1 focuses in on our period of study, 2016 to 2021, and displays the proportion of cases the Court decided each year that involved each of the fourteen specific issue areas in the Supreme Court Database (Reference Spaeth, Epstein and MartinSpaeth et al., 2021). Some areas are more variable than others, but civil rights, criminal procedure, and economic activity predominate, with each making up the largest proportion of cases in at least one year. Judicial power is the only other issue area that constitutes the highest proportion in any given year (2017), but the number of cases in this area is much more variable.

New justices not only bring preferences over issues areas to the bench, they also bring preferences over the policy outcomes they would like to see in the cases they hear (Reference Segal and SpaethSegal & Spaeth, 2002) and the legal reasoning that should be used (Reference LiebellLiebell, 2023). Using coding from the Supreme Court Database, Figure 2 shows trends in the ideological direction of the Court’s decisions over time in all cases (left panel) and only nonunanimous cases (right panel).Footnote ¹⁷ For most of the last four decades, the Court has made a greater proportion of conservative decisions (between approximately 50 and 60%, depending on the year) than liberal ones (between 40 and 50%). When the cases in which all members of the Court agreed are removed, we see that the rate of conservative decisions is generally higher, and has leaped to nearly 70% of nonunanimous cases since Trump’s third appointee, Amy Coney Barrett, took the bench.

Figure 2 The ideological direction of SCOTUS cases over time, 1980 to 2021. Left panel reflects all cases and right panel reflects only the nonunanimous cases

This is consistent with the make-up of the Court and other research that finds the Court is increasingly more conservative than the public (Reference Jessee, Malhotra and SenJessee, Malhotra, & Sen, 2022). In every term since 1970, the Court has had at least five justices appointed by Republican presidents, and every chief justice during the same period has been appointed by a Republican. Given this consistent conservative majority, it is perhaps more surprising that there have been brief periods, most notably in the early 2010s, in which liberal decisions outpaced conservative ones.

However, looking at trends for all cases can be misleading. As noted earlier, the issues on the Court’s docket vary from year to year. Even when cases are in the same issue area, some cases are closer calls than others, making year-to-year comparisons imprecise (Reference BaumBaum, 1988; Reference Clark and LauderdaleClark & Lauderdale, 2010). Experts suggest that reversals are more accurate reflections of the Court’s ideology than affirmances because the Court most often grants review to reverse a lower court decision with which it disagrees. Affirmances, then, represent a miscalculation on behalf of the litigants, the justices, or both, about how the case aligns with the majority of justices’ preferences (Reference McGuire, Vanberg, Smith Jr and CaldeiraMcGuire et al., 2009).

Figure 3 displays the percent of affirmances and reversals by ideological direction during the time period of our study, 2016 to 2021.Footnote ¹⁸ Interestingly, we do not see the differences Reference McGuire, Vanberg, Smith Jr and CaldeiraMcGuire et al. (2009) documented between affirmances and reversals during this period. Instead, affirmances actually seem more conservative, particularly after Justice Kavanaugh’s confirmation. This could indicate a change in the justices’ behavior, in which they are granting more appeals of lower court decisions they want to affirm and make national precedents. It could also be a result of the way liberal and conservative decisions are coded in the Supreme Court Database, which regards decisions favoring plaintiffs alleging violations of their rights to free exercise of religion or free expression as liberal decisions. In prior periods, these claims often involved small, unpopular groups, but in the Court’s recent decisions, they have tended to involve mainstream, conservative religious groups – for example, a Christian coach who prayed on a public school football field,Footnote ¹⁹ a Christian organization that wanted to fly their flag in front of Boston City Hall,Footnote ²⁰ and parents who wanted the state to pay for tuition for their children to attend Christian religious schools.Footnote ²¹ These decisions – all of which were decided in the 2021–2022 term and resulted in Supreme Court victories for the religious plaintiffs – confound traditional definitions of liberal decisions by intermixing issues of free exercise, free expression, and the Establishment Clause and favoring majority rather than minority interests.Footnote ²²

Figure 3 The ideological direction of SCOTUS cases over time, 2016 to 2021. This graph splits by case disposition, whether they are affirmances (left) or reversals (right). Cases with other dispositions were excluded

Whatever flaws may exist in our current measures of the ideology of Court decisions, it seems clear that the Court has become more conservative since Trump’s second nominee, Kavanaugh, was confirmed. Equally significant, the way the Court explains its reasoning – which is important for how lower courts interpret the law moving forward – has shifted as well. The legal reasoning in recent cases like New York State Rifle & Pistol Association, Inc. v. Bruen,Footnote ²³ Kennedy v. Bremerton School District,Footnote ²⁴ and Dobbs v. Jackson Women’s Health OrganizationFootnote ²⁵ is a particular form of originalism that will shape decision-making throughout the federal judiciary for decades to come (Reference Blocher and WillingerBlocher & Willinger, 2023; Reference Wilson and Hollis-BruskyWilson & Hollis-Brusky, 2023; Reference ZieglerZiegler, 2023).

Changes to the Supreme Court’s issue agenda, reasoning, and decisions are substantively important consequences of changes in the Court’s membership. They shape the rights and remedies available to all US residents in ways well illustrated by the cases just cited. Those cases limited states’ ability to restrict gun ownership, approved sectarian prayer at a public school event, and overturned half a century of precedent protecting the right to abortion care. In addition to these serious implications, changes in the Court’s membership may also shape the strategic incentives for litigants and amicus curiae to turn (or not to turn) to the Court for relief. This has important implications for representation in government and policymaking.

1.3 Representation and Participation at the Supreme Court

Policy entrepreneurs – those interested in changing the law in specific policy areas – watch the Court carefully and bring cases that will give the justices opportunities to make policy they favor (Reference BairdBaird, 2007). Skilled attorneys tailor their arguments to the current members of the Court (Reference Drolc, Merrill and SchoenherrDrolc, Merrill, & Schoenherr, 2023; Reference WedekingWedeking, 2010) and coordinate amicus curiae who can bolster their arguments (Reference Larsen and DevinsLarsen & Devins, 2016). Interviews with attorneys who practice before the Supreme Court reinforce the view that litigants and their attorneys are strategic actors who think carefully about whether to bring a case to the Court or to get involved in an amicus brief (Reference Hazelton and HinkleHazelton & Hinkle, 2022).

This strategic calculus implies that there will be times when those with potential claims and policy goals choose to opt out of participation in Supreme Court litigation. This Element explores whether and how the membership of the Court impacts the decision to ask the Supreme Court to review a case or to provide it with an amicus brief. This is important to understand because those who choose not to submit petitions for certiorari or participate in amicus briefs are also shaping the Court’s docket and opinions by their absence. The Court does not start from scratch in deciding what issues to review; it is dependent on litigants to bring it cases. Thus, the petitions and briefs the Court receives (and the absence of those it does not receive) shape the Court’s docket as well as case outcomes and opinion language (Reference Collins, Corley and HamnerCollins, Corley, & Hamner, 2015; Reference Hazelton and HinkleHazelton and Hinkle, 2022; Reference Hazelton, Hinkle and SpriggsHazelton, Hinkle, & Spriggs, 2019).

Judicial advocacy is just one tool in the advocacy toolkit of those seeking to influence public policy (Reference GrossmannGrossmann, 2012; Reference WalkerWalker, 1991). Under our constitutional system, the legislative branch plays the central role in policymaking, and most groups that are active in policy advocacy devote at least some of their efforts to legislative advocacy (Reference GrossmannGrossmann, 2012). Court advocacy, such as filing lawsuits or otherwise engaging in litigation, is generally the least frequent activity, in part because opportunities for participation are more structured and formal and require legal expertise (Reference Nownes and FreemanNownes & Freeman, 1998; Reference Schlozman and TierneySchlozman & Tierney, 1986). It is also because the odds of making substantial policy change through the courts are low (Reference RosenbergRosenberg, 2008). When parties win in the lower courts, their cases usually have limited policy effects. And because the Supreme Court takes so few cases – it grants review of approximately 1% of the cases it is asked to review each year (representing only 0.02% of the cases that are filed in the district courts annually) – the chances of any one case having national impact are very small. Individuals, as well as organized interests, need to be strategic in deciding whether to litigate because litigation is an expensive and time-consuming process, and even the most experienced attorneys can provide no guarantee of success (Reference McCannMcCann, 2006). Moreover, financial costs are not all that are at stake when litigating. While a party who seeks legislation and fails usually just ends up with the status quo, a party who seeks policy change at the Supreme Court could end up with national precedent that goes against their interests. For example, the advocates who sought to have Mississippi’s restrictive abortion law struck down ended up losing national constitutional protection for abortion rights in Dobbs v. Jackson Women’s Health Organization (2022).

Groups develop their expectations about the Court’s receptiveness to their claims in several ways. First, as noted above, Supreme Court confirmation hearings are highly salient, public affairs. In televised hearings, senators question the nominees about a wide range of topics, from their judicial philosophy to their views on particular issues (Reference Collins Jr, Ringhand and BoydCollins, Ringhand, & Boyd, 2023). If the nominee has been a judge before, their prior decisions provide additional insights into their judicial philosophy and preferences (Reference HittHitt, 2013). While litigants themselves might not be familiar with these materials, their lawyers – if they are represented – have access to this information and likely use it to shape their recommendations. Finally, the longer a justice sits on the bench, the more evident their preferences over issues and methods of interpretation become through their opinions and votes in cases. These sources of information provide litigants a foundation on which to base strategic decisions.

Litigants who receive these signals and do not expect to win may decide to avoid the financial and policy costs of filing for review in the first place. If an organized interest suspects that the courts will not be receptive to its position, it can shift its resources and effort to another policymaking venue, just as the abortion rights groups have done with their shift to a focus on state-level advocacy. Of course, the opposite is also true. Groups that see the courts leaning in their direction may also shift focus toward the courts. In sum, groups turn to the courts in a strategic manner that may shift as the composition of the courts’ membership evolves.

We expect that the increasing conservatism of the Supreme Court that resulted from President Trump’s appointments shifted the strategic incentives of different potential parties and amici to participate in Supreme Court litigation. We expect these effects to be cumulative, with the strongest shifts occurring after the more solidly conservative Brett Kavanaugh replaced frequent swing justice Anthony Kennedy, and then again after conservative Justice Amy Coney Barrett replaced the liberal icon, Ruth Bader Ginsberg, creating a 6-3 conservative supermajority on the Court. Drawing on decades of Supreme Court literature, we expect to see changes among three main types of groups: those who are politically disadvantaged, those who are well-resourced repeat players, and those who are pursuing ideological agendas. We expect some of these groups to decrease participation before the Court and others to increase participation, as explained below.

1.4 Data and Measures

To test our expectations, we assembled two original datasets on participation in Supreme Court litigation. The first is a dataset of all petitions for certiorari on which the Supreme Court took some kind of action in its October 2016 through October 2021 terms.Footnote ³⁰ We pulled the docket numbers, party names, and actions from the Supreme Court Journals, the official minutes of the Court.Footnote ³¹ The vast majority of the petitions for certiorari were denied. Approximately 1% were granted full hearing, and the remainder were granted and summarily decided. During our period of study, the Court denied between 4,722 and 6,157 petitions per year and granted full review to between 60 and 78 petitions per year.Footnote ³² The journals include the dates the Court took action rather than the dates the petitions for certiorari were filed, so we used Python to scrape the filing dates from the Supreme Court website.Footnote ³³ The second dataset includes all parties and amicus brief signers on the 396 cases the Court gave full hearing during the 2016 through 2021 terms.Footnote ³⁴ This data set was collected using SCOTUSblog’s term archive, which provides information on all cases heard each term since 2007, with links to all case filings.Footnote ³⁵

We connect information on the parties and amicus participants with three temporal events: the confirmation of Neil Gorsuch on April 7, 2017, the confirmation of Brett Kavanaugh on October 6, 2018, and the confirmation of Amy Coney Barrett on October 26, 2020. We chose confirmation date and not nomination date because litigants and amici cannot be certain when or whether a nominee will be confirmed, and thus whether the nominee will be involved in deciding their cases, injecting uncertainty into their strategic calculus. Under Supreme Court Rule 13, litigants have only 90 days after a final decision by the lower court to petition for certiorari, so they cannot wait around to see how a nomination plays out before deciding.Footnote ³⁶ Further, cases heard after nomination but before confirmation for both Justices Gorsuch and Kavanaugh were heard by a Court split evenly between liberal and conservative appointees. In our data, 74 cases were heard prior to any of the Trump justices’ confirmations, 96 were heard between the confirmations of Gorsuch and Kavanaugh, 129 were heard after the confirmations of both Gorsuch and Kavanaugh but before Barrett’s, and 97 were heard after all three were confirmed. We then compare filings in the four distinct time periods in our data: before any Trump appointees took the bench, after Gorsuch but before Kavanaugh and Barrett, after Kavanaugh’s confirmation but before Barrett, and after all three.

Next, we categorize the parties and amici to test our hypotheses. Recall that Hypothesis 1 suggested that the most politically disadvantaged groups will become less likely to submit petitions for certiorari or amicus briefs as the Trump appointees join the bench. To operationalize political disadvantage, we consider both the social construction and the power of those involved. Social construction refers to whether a group is considered “good” or “deserving” of favorable policy outcomes (Reference Schneider and IngramSchneider & Ingram, 1993). Power here refers to political power: the ability to make campaign contributions or vote in significant numbers, for example. A group can be seen as undeserving but still powerful, like big businesses, which are often perceived as greedy and exploitative, but play an important role in the national economy that may command policymakers’ attention. They also have substantial resource advantages that can be used to meet their political goals. Conversely, groups can be seen as deserving but not powerful. For example, children are generally valued by society but lack the right to vote or the ability to make financial contributions, which reduces their influence with policymakers (Reference WidnerWidner, 2020). Finally, some groups are both negatively constructed and not powerful. An example of this is people who are incarcerated. They are often viewed as wrong-doers who need to be punished, and many of them have lost the right to vote due to felony convictions. While both negative construction and low political power are disadvantages in the policymaking process (Reference Schneider and IngramSchneider & Ingram, 1993), those who have both negative constructions and low power are the most politically disadvantaged.

To operationalize disadvantage, we use crowdsourced ratings of social construction and political power from Reference Kreitzer and SmithKreitzer and Smith (2018). Because the social construction of a group is inherently about public perception, crowdsourcing provides a powerful way to assess how different groups are viewed. Reference Kreitzer and SmithKreitzer and Smith (2018) asked respondents to separately rate a group’s “deservingness” and their power. They found that some groups’ construction varied greatly with the respondents’ political affiliations, while other groups’ constructions were more consistent. Though Reference Kreitzer and SmithKreitzer and Smith (2018) consider more than seventy distinct groups, here we focus on only the most disadvantaged as a conservative test of our expectations. We coded only groups who were at least one standard deviation below the mean rating in deservingness (we refer to this as negative construction), power, or both. Table 2 displays the groups that were coded in each category. Because Reference Kreitzer and SmithKreitzer and Smith (2018) did not include every possible litigant group in their data, this list is likely underinclusive of the most disadvantaged populations, but it provides a useful starting point for analysis.

We coded all petitioners and amicus filers with two dummy variables: (1) Negative Construction, which is coded as 1 if the group is listed in Table 2 as having negative construction (or low deservingness) and 0 otherwise, and (2) Low Power which is coded as 1 if they are listed as having low power and 0 otherwise. A petitioner or amicus filer did not need to be an individual member of a group to be coded as 1; interest groups that advocate for the interests of the group were coded as 1 as well. For example, public defenders’ offices and associations of criminal defense lawyers generally file amicus briefs making arguments for the criminal defendant’s side of a case, so they are coded as 1 for both negative construction and low power. Trade associations that represent big banks or big corporations are coded as negative construction but not low power. If an organization represents many groups, some of which have a negative construction or low power and some of which do not, they were coded as 0. For example, chambers of commerce, which represent both large corporations and small businesses, were not coded as having a negative construction. Similarly, the American Civil Liberties Union, which sometimes represents criminal defendants and undocumented immigrants but also represents many other groups, was coded as neither negative construction nor low power. Again, this approach is likely under-inclusive, but it allows for a conservative test of our expectations. For businesses, we coded big banks and corporations and media as falling under negative construction. To operationalize this, we coded all financial services and media companies as negative construction. For other businesses, we only coded them as negative construction if they were on the Fortune 1000 list during our period of study. All other businesses were coded as 0 on both negative construction and low power. This conservative coding was chosen to limit subjective judgments. We use our two dummy variables to classify the petitioners and amicus participants into four mutually exclusive categories: negative construction only (negative construction but not low power), low power only (low power but not negative construction), negative construction and low power, and not disadvantaged (neither negative construction nor low power).

For our repeat-player analysis, we follow scholars such as Reference Songer and SheehanSonger and Sheehan (1992) and Reference Hazelton and HinkleHazelton and Hinkle (2022) and classify our petitioners and amicus filers into the following mutually exclusive categories: business, government, individual, or interest group. For governments, we also code whether they are federal, state, international, tribal, or local government actors. These distinctions are predictive of variation in case participation. Reference Hazelton and HinkleHazelton and Hinkle (2022), for example, find that in cases where the Court gave a full hearing over a thirty-year period, only 25% of individuals who filed a brief had done so previously, compared to 48% for businesses, 56% for federal government actors, and 72% of interest groups and subnational government filers. This coding also provides a proxy for resources for litigation (Reference Songer, Sheehan and HaireSonger et al., 1999). Businesses generally have resource advantages over individuals, while governments, particularly the federal government, tend to have the greatest resource advantages.

Finally, to test our ideological hypotheses, we code both state governments and interest groups into ideological categories. Specifically, we code state government actorsFootnote ³⁷ into three categories: blue, red, or purple, based on whether the state’s attorney general and governor are both Democratic (blue), both Republican (red), or belong to different parties (purple). We chose to focus on attorneys general and governors for a few reasons. First, governors, as the head of their state, are considered to be chief legislators and party leaders, analogous to presidents and other chief executives with significant power (Reference Kousser and PhillipsKousser & Phillips, 2012). Second, because states’ attorneys general are elected officials in most states and they represent their states in litigation, they are key players in litigation before the Court.Footnote ³⁸ Governors are also often named in litigation as the executives of their states, so they also bear a closer relationship to the court process than does the state legislature or state supreme court, for example.

Similarly, we coded interest groups into three mutually exclusive categories: liberal, conservative, and nonpartisan. Liberal groups are those that explicitly mention a liberal or progressive agenda on their websites or other public materialsFootnote ³⁹ (e.g., Public Citizen or Texas Progressive Action Network), and conservative groups are those that mention conservative or family values (e.g., Washington Legal Foundation or Conservative Legal Defense and Education Fund). All those that did not expressly proclaim a liberal or conservative mission were labeled as nonpartisan. Groups labeled nonpartisan may pursue policy objectives more associated with one party or ideology – for example, reproductive rights or gun rights – but do not frame their missions in ideological terms. Most interest groups are not explicitly partisan: we identified only 170 liberal and 385 conservative organizations among the more than 6,000 interest groups in our data.

1.5 Looking Ahead

Up to this point, we have discussed the strategic incentives of litigants and amici together. However, the decision to file a petition for certiorari is not the same as the decision to participate in an amicus brief. The specific outcome of the case is usually more important to the parties, whereas amici are more often interested in the legal and policy aspects of a decision (Reference Hazelton and HinkleHazelton & Hinkle, 2022). Further, the costs of litigation fall more fully on the parties than on amici. Petitioners generally have to shoulder their own costs of litigation. In contrast, individuals and groups with shared interests typically co-sign amicus briefs, spreading out the work and the cost (Reference Box-Steffensmeier, Christenson and HittBox-Steffensmeier, Christenson, & Hitt, 2013). Other differences between petitioners and amici necessitate different empirical approaches. For example, coding petitioners for certiorari into the categories described earlier is substantially more challenging than coding amicus participants. Petitioners are more numerous and are much more likely to be individuals. Those who sign amicus briefs are more likely to be repeat players and to be well known. For all these reasons we divide our analysis in the sections to come.

Section 2 takes a detailed look at the petitioners for certiorari. It discusses the techniques we employed to code them into our categories of interest, describes what we have learned about who filed petitions during the years of our study, and analyzes whether and how that has changed as Trump’s appointees have been confirmed. We find support for our political disadvantage and ideological hypotheses, but less support for our repeat player hypotheses. Trends suggest that the most politically disadvantaged groups have been filing fewer and fewer petitions as each Trump justice joined the bench, while Republican-controlled states and explicitly conservative interest groups have been ramping up their efforts to bring cases to the Court. Most other categories of filers have remained fairly steady in their rates of filing or seen only modest increases or decreases.

Section 3 shines a similar light on amicus participants. It explains two different opportunities amici have to file briefs – the certiorari stage and the merits stage – and tests our hypotheses at each stage. Our most significant findings relate to our ideological expectations. Red states and explicitly conservative interest groups have dramatically increased their participation in amicus briefs at both stages as each Trump justice has been confirmed. We also find the resource typology for haves and have-nots and repeat players is in some ways a poor fit for studying amici. The individuals who participate in amicus briefs, at least in our period of study, are sophisticated political, legal, and academic elites, and the have-nots are most commonly represented by skilled advocacy organizations.

Finally, Section 4 reflects on the major findings of this study. It argues that who is on the Court influences political participation and representation in ways that are worthy of our attention. We also look ahead to the potential impact of President Joe Biden’s nominations and those of administrations to come. Changes in the perceived and actual receptiveness of one of our three branches of government to different groups have implications for law, policy, and political inclusion. These conclusions also reach far beyond the courts, and we argue that the importance of court appointments should be made more visible to litigants, voters, and candidates for offices who can affect the process.

2.1 Coding Petitioner Types

The greatest challenge for coding the petitioners into the categories described in Section 1 is the sheer volume of them: over our time period, more than 33,000 petitions for certiorari were acted on by the Court, and almost 30,000 of the petitioners filed only one petition. To test our expectations, we needed to code these petitioners into the political disadvantage, repeat player, and ideological frameworks described in Section 1. Fully hand-coding the sample was not realistic, so instead we used a combination of hand-coding, text analysis techniques, and probability matching. We first cleanedFootnote ⁴⁴ the data for ease of matching, then assigned petitioners into our categories of interest using the following steps:

1. 1. Hand-Coded Amicus Curiae. We began by hand-coding all the amicus participants in the dataset used for Section 3. This was the logical starting place because there are fewer unique individuals and organizations in that dataset, and more of them are easily recognizable. We then matched the amicus coding to the petitioner dataset, allowing us to code any individual, government, or organization that appears in both datasets.

   • Example: 3M Company was in our amicus curiae dataset and is also a petitioner in a case denied cert. We coded 3M as a business and as 1 on negative construction and 0 on low power because they are in the Fortune 1000.

2. 2. Hand-Coded Parties in Cases Granted Cert. We also hand-coded the parties in every case in which the Supreme Court granted cert and in which argument was heard during the October 2016 to 2021 terms, and matched these to the petitioner dataset.

   • Example: Bank of America is a party in a case granted cert and is also a petitioner in a case denied cert. We coded Bank of America as a business, and as 1 on negative construction and 0 on low power because they are a big bank and are in the Fortune 1000.

3. 3. Hand-Coded Petitioners. We then hand-coded over 6,000 randomly selected petitioners with the help of research assistants.Footnote ⁴⁵

   • Example: Aaron Brothers, as Independent Executor of the Estate of William Slade Sullivan, was hand-coded as an individual and as 0 on both negative construction and low power because he does not fit any of our disadvantage categories.

4. 4. Respondent Dictionary. The identity of the respondent can also provide information on the identity of the petitioner. We took the full case names of the remaining petitions and coded based on common phrases or words in the respondent names.Footnote ⁴⁶

   • If the respondent is a warden, jail, or correctional facility, the petitioner is a prisoner, and therefore coded as individual with negative social construction and low power.

   • Example: Mickey Frank Pryor was labeled as a prisoner in our data as he was suing Lorie Davis, Director of the Texas Department of Criminal Justice Correctional Institutions Division. We code Pryor as an individual and as 1 on both negative construction and low power.

5. 5. Petitioner Dictionary. We then created a dictionary with nearly 200 terms to capture common phrases that would correspond to our categories in the petitioners’ data and used these for additional coding.Footnote ⁴⁷

   • If the petitioner’s name included the term ‘sheriff’ they would be coded as a government with neither low power and negative construction.

   • Example: Gannett Co., Inc., is edited to be Gannett Company, Incorporated. Because the petitioner name includes both company and incorporated, it is identified as a business with 0 on low power.

6. 6. IFP Coding. The docket numbers for paid petitions and in forma pauperis (IFP) petitions are numbered in different sequences.Footnote ⁴⁸ Any case in which the portion of the docket number after the hyphen is 5001 or higher (i.e., 16-5001), involves an IFP petitioner and was therefore coded as an individual.

   • Example: Aaron Matthew Oleston filed under docket number 21-7442, so he was labeled as an individual.

7. 7. Fuzzy Matching. We used fuzzy matching for the remaining petitioners missing their repeat player categories, with a maximum Jaro-Winkler (JW) distance of 0.2.Footnote ⁴⁹ We chose not to use fuzzy matching for negative construction and low power as these are much more context dependent, especially for individuals in our data.

   • We model this approach after Reference Van der LooVan der Loo (2014), which describes how to deploy approximate string matching in R. The JW distance “measures the number of matching characters between two strings that are not too many positions apart and adds a penalty for matching characters that are transposed” (p. 119). For our purposes, it is less important that these names match exactly, but instead that the method correctly identifies the petitioner type.

   • Example, Aaron Jensen was matched to Aaron J Webster with a JW distance of 0.16666667, so Jensen was labeled as an individual. These names are not identical and the petitioners are not the same person, but JW distance does well on short strings like names, which is sufficient for our broad coding of repeat player types (Reference Van der LooVan der Loo, 2014). We also performed spot checks on the data to check the success of the fuzzy matches.

   • This approach allows us to easily identify thousands of individuals in our data without hand-coding them. There are important tradeoffs in this decision, however. Hand-coding each observation would not be possible given the large quantity of observations: one alternative is that we could rely on a lower number of hand-coded observations. Or, we could employ the strategy we use here and use fuzzy matching with potential room for error, but be able to use the full dataset in our results (Reference Kaufman and KlevsKaufman & Klevs, 2022). Because we seek to describe the overall pattern of litigant participation, rather than just a subset, we use fuzzy matching, but acknowledge room for noise in this approach. We do also note, though, that we performed extensive spot checks in our iterative fuzzy matching process to ensure there are not obvious errors.

8. 8. Additional Hand-Coding. At this point, all of our petitioners in the dataset had been coded into their repeat player types. We then hand-coded the political disadvantage variables (low power and negative construction) for all businesses, interest groups, and governments, which combined make up about 8% of all petitioners in our period of study. This also allowed us the opportunity to double check and correct any errors from the fuzzy matching.

   • For example, In-N-Out Burger, Incorporated was labeled as a business using the dictionary approach in Step 5. We then hand-coded it as 0 on negative construction and low power, because while it is a business, it is not a “big business” under our definition (see Section 1).

9. 9. Proportion Matching. Individuals make up 92% of all petitioners. After all of the previous steps, about 55% of them were still not coded for our negative construction and low power variables. We used proportion matching to assign the remaining political disadvantage coding for those individuals. Specifically, we pulled an additional random sample of 500 of these individuals and hand-coded them for negative construction and low power. We then split the sample by IFP status, because IFP petitioners are far more likely to be politically disadvantaged. Per our random sample, IFP petitioners had an 80% chance of being low power and negative construction. Petitioners who were not IFP had a 28% chance of being low power and a 25% chance of being negative construction. We used these proportions to randomly assign coding on these two variables to the remaining individuals in our data based on IFP status.

   • This approach, while not perfect, allows us to see the full picture of participation by individuals in our data. One alternative to proportion matching is iterative hand-coding. Not only would this be prohibitively time-consuming, but also may be impossible. There may be zero records (legal or otherwise) of these individuals online to determine their identity, let alone their traits to determine negative construction and low power status. Therefore, while the proportion matching is not exact, we are reasonably confident that it mirrors the approximate proportions of individuals in our data by political disadvantage.

   • We checked for whether the distribution of these variables changes over our four discrete time periods in the Appendix; they do not. See Appendix Section B4 for these comparisons.

We use these codings for our petitioner analyses in this section. To verify our assumptions regarding the proportion matching, we conducted similar analyses using only samples from the hand-coded petitioners. These are presented in the Section B5 of the Appendix. The trends are substantially similar.

2.2 Analysis

To test the hypotheses described in Section 1, we use time trends to see whether and if the identities of petitioners for certiorari changed over time as Trump’s appointees joined the Supreme Court. The analyses that follow use the number of filings by relevant group type by month to plot these trends.Footnote ⁵⁰ While this relatively simple analysis does not allow us to claim that the changing make-up of the Supreme Court caused the trends we are seeing, it does provide evidence that the composition of the parties petitioning the Court for review has changed over time.

2.2.1 Trends for Politically Disadvantaged Petitioners

As we discussed in Section 1, courts have long been seen as the most receptive policymaking venue for politically disadvantaged groups (Reference CortnerCortner, 1968; Reference VoseVose, 1959). However, we expect that the make-up of the courts generally and the Supreme Court more specifically may shift how receptive disadvantaged petitioners expect the Court to be to their interests. We expect that the most politically disadvantaged will become cumulatively less likely to seek Supreme Court review as each new Trump nominee takes the bench. We conceptualize the most politically disadvantaged groups as those that have both negative social construction and low power (Reference Kreitzer and SmithKreitzer & Smith, 2018; Reference Schneider and IngramSchneider & Ingram, 1993).

Figure 4 shows the trends in numbers of petitions for certiorari filed per month over time for each of our four mutually exclusive groups: those with low power only, those with negative construction only, those with both negative construction and low power (the most disadvantaged), and those who are not disadvantaged under our conservative definition. Petitioners with only one form of disadvantage – either low power or negative construction – file at remarkably similar rates that stay relatively consistent over time. Petitioners that are not disadvantaged file at a higher rate with slightly more fluctuation. On the other hand, the difference between the most disadvantaged petitioners and the other categories is stark. While the number of petitions filed by most group types stays more or less flat, petitions from the most disadvantaged litigants decline dramatically. The drop begins before the end of 2018, drops most after Justice Kavanaugh’s confirmation, and continues dropping after Barrett’s appointment, before leveling off shortly thereafter.

Figure 4 Monthly cert filings by political disadvantage pre- and post-candidate confirmation with a plotted loess line

One potentially confounding factor is that our period of study overlaps with the COVID-19 pandemic and associated national emergency. Courts and related institutions like jails and prisons faced challenges in safely conducting their work during this time (Reference DolovichDolovich, 2020). The Bureau of Justice Statistics found that the nation’s prison population decreased by more than 16% in the first year of the pandemic. This, along with restrictive safety practices adopted in many jail and prisons during this period, may help account for the falling number of petitions by the most disadvantaged litigants, many of whom are criminal defendants and incarcerated people.Footnote ⁵¹ However, there is reason to believe that the trend in filings by the most disadvantaged petitioners was not predominantly driven by the pandemic. First, the decline in disadvantaged petitions begins during the 2018 term – more than a year before the pandemic. Second, petitions in other categories do not show any notable sustained decline around the time of the pandemic, suggesting that pandemic-related delays in lower court processing of cases were not driving changes in petitioner composition; otherwise, all categories of filers would be expected to have fallen.

While it is true that the number of individuals experiencing incarceration has declined over this period, that decrease is not enough to explain the sharp decline in petitions. According to data collected by the US Department of Justice, prison populations have declined approximately 14% over our period of study.Footnote ⁵² In contrast, the number of petitions for certiorari from the most politically disadvantaged petitioners fell by over 30%. Although COVID and decarceration may have played some role, the overall trend is consistent with our first hypothesis.

2.2.2 Trends for Repeat Players

In Section 1, we explained that we expected an increasingly conservative Supreme Court to exacerbate the advantage of repeat players with greater resources. We expected these repeat players – specifically businesses, interest groups, and governments – to react strategically to changes in the Court’s composition by increasing their advocacy before the Supreme Court. Conversely, we hypothesized that individuals would expect the Court to be less receptive as the Court grew more conservative and would become less likely to file. However, at the beginning of this section we also speculated that IFP petitioners – all of whom are individuals and who may be relying on the courts for literal life-and-death decisions – might be less likely to be affected by strategic incentives and less responsive to changes in the Court.

The trends in Figure 5 are inconsistent with our expectations. Filings by most group types stay relatively flat. Business filings decrease slightly, and IFP individuals decrease most dramatically. This is consistent with the political disadvantage analysis, but inconsistent with our hypothesis concerning behavior of repeat players.

Figure 5 Monthly cert filings by repeat player filer type pre- and post-candidate confirmation with a plotted loess line

While overall government filings look mostly flat in Figure 5, when we disaggregate by government type we see a more nuanced picture. As Figure 6 shows, filings by state governments were on the decline when Gorsuch was confirmed, but began to increase again after Kavanaugh’s confirmation, then skyrocketed after Barrett’s confirmation. By the end of 2021, they were nearly double what they were in 2016. Filings by the federal government increased after Kavanaugh’s confirmation in a period during which both the Court and the administration were conservative, but then declined in 2021 as President Biden took office. This is consistent with our conditional expectations embodied in Hypotheses 2, 4, and 5. Filings by international, local, and tribal governments do not follow any clear pattern.

Figure 6 Monthly cert filings by government actors pre- and post-candidate confirmation with a plotted loess line

2.2.3 Trends by Ideology of Filers

Figure 7 shows that the increase in filings by state governments was not a uniform increase in participation by all state governments. Instead, the increase is driven by petitions for certiorari from red (Republican-controlled) states. As the Court has grown increasingly conservative, red states have filed more petitions for certiorari. Here, the biggest spikes come after Justice Barrett was confirmed, giving the Court a solid 6-3 conservative majority. This is consistent with Hypothesis 4. In contrast, purple (mixed control) states filed very few petitions at any point during this period. Blue (Democratic-controlled) state petitions are only slightly higher than those filed by purple states, though they show a brief bump in the first part of 2020. This does not support the decline predicted by Hypothesis 5, but it does offer more general support for our underlying theoretical expectation that liberal and conservative governments will respond differently to the Court’s membership.

Figure 7 Monthly cert filings by state government actors pre- and post-candidate confirmation with a plotted loess line

Red states are not the only conservative actors who increased their presence before the Court. Figure 8 shows that explicitly conservative interest groups have increased the number of petitions for certiorari they file with the Court since Trump’s nominees were confirmed. At the beginning of our period of study, there are very few petitions from either explicitly conservative or explicitly liberal interest groups. Filings by conservative groups began to increase slightly before Justice Kavanaugh was confirmed, grew slightly until Justice Barrett was confirmed, and then increased again after her confirmation. This is consistent with Hypothesis 4 and our expectation that effects of the Court’s membership changes will be cumulative. Nonpartisan interest groups still file more petitions overall and have increased their activity somewhat, but the gap between nonpartisan and conservative interest groups is narrowing. We do not see a decrease in filings by expressly liberal interest groups in support of Hypothesis 5, but that may be because they were already essentially zero.

Figure 8 Cert filings by interest group type pre- and post-candidate confirmation with a plotted loess line

3.1 Amicus Participants and the Changing Make-up of the Court

Although change in the membership of the Court does not, by itself, have a significant effect on the number of amicus briefs filed, scholars suggest that ideological heterogeneity on the Court results in more briefs being filed because there is more uncertainty about the potential outcome (Reference Salzman, Williams and CalvinSalzman, Williams, & Calvin, 2011). Conversely, we might expect that the decreasing ideological heterogeneity on the Supreme Court with the arrival of the Trump appointees will depress amicus filings, at least by those who do not expect the Court to be receptive to their views. Further, groups’ decisions to participate as amicus curiae are often spurred from the outside. Members of the Supreme Court bar frequently identify organizations that should be heard from and help to shape the messages amici present (Reference Larsen and DevinsLarsen & Devins, 2016). These lawyers are keenly attuned to who is on the Supreme Court and what arguments are most effective in reaching them. Therefore, they would likely want to shape the pool of potential amicus participants to maximize the likelihood of influence.

Amicus briefs are filed at two distinct stages of Supreme Court review. Because the Court has discretionary jurisdiction over almost all of its docket, litigants seeking Supreme Court review must first file a petition for certiorari, asking the Court to grant review of their case. Amicus briefs are often filed at this stage to encourage or discourage the Court from hearing the case at all. Evidence suggests that these briefs provide an important signal to the Court about the importance of a case. The more briefs filed in a case at the certiorari stage – no matter whether those briefs are advocating grant or denial – the more likely the Court is to grant review (Reference Caldeira and WrightCaldeira & Wright, 1988; Reference Schoenherr and BlackSchoenherr & Black, 2019), though the value of this signal may have decreased somewhat as amicus participation at the certiorari stage has become more common (Reference Caldeira, Wright and ZornCaldeira, Wright, & Zorn, 2012). Amicus briefs can also be filed after the Supreme Court has granted review, to aid it in its consideration of the merits of the case. In other words, briefs filed at the certiorari stage seek to influence the Court’s agenda, while those filed at the merits stage seek to influence outcomes and policy. The decision to file an amicus brief is always strategic, though different factors may shape strategic decision-making at each stage. Specifically, groups that expect the Supreme Court to be sympathetic to their interests should be more active at the certiorari stage than groups that do not. Once the Court has already decided to hear a case, incentives shift for groups who do not expect a favorable reception. Even if they would not have wanted a case to be heard, groups may feel compelled to offer a brief at the merits stage in order to counteract arguments that harm their interests or to show their supporters that they are working to defend their policy preferences (Reference HansfordHansford, 2004b).

3.2 Data

To analyze shifts in amicus behavior resulting from Trump’s nominees to the Supreme Court, we collected information on the 396 cases heard in the 2016 to 2021 termsFootnote ⁵³ and gathered all the amicus signers on those cases – in total, over 53,000 signers on nearly 6,000 individual briefs (Reference Gunderson, Widner and MacdonaldGunderson, Widner, & Macdonald, 2023). Figure 9 displays the number of amicus briefs and amicus signers for cases heard between 2016 and 2021. This figure includes briefs at the certiorari stage (left panel) and merits stage (right panel). The average case had over 13 amicus briefs filed (with a median of 9), with over 130 organizations, individuals, or businesses signing on as amicus participants on average (with a median of 41). Some cases had no amicus briefs and some had more than 100. The maximum number of signers on one case was over 7,000. On average, from 2016 through 2021, there were approximately 1,000 amicus briefs filed in decided cases each year, with anywhere from 5,000 to 14,000 participants on those briefs. The number of briefs has trended up slightly over the time frame of our study – rising from an average of 12 briefs per case in 2016 to 15 briefs per case in 2021 – while the number of participants on those briefs has exploded.

Figure 9 Amicus briefs and amicus participants over time at the cert stage (left panel) and merits stage (right panel), 2016 to 2021

Once we disaggregate this information by stage, however, it is clear that most amicus activity is occurring at the merits stage. Cert stage amicus participants annually typically number in the hundreds, but merits participants regularly exceed five thousand. Between 100 and 200 amicus briefs are filed at the cert stage per year, compared to nearly a thousand at the merits stage. Interestingly, Figure 9 also reflects fairly consistent growth in cert filings over time, but more variation in merits amicus participants per year.

We connected our dataset of amicus participation with the confirmation dates of Justices Gorsuch, Kavanaugh, and Barrett. Rather than pooling amicus briefs across both stages at which they could be filed – the cert or merits stages – we look at each stage separately because of the potential for different strategic incentives at each stage. We then hand-coded all of our political disadvantage, repeat player, and ideological variables for all amicus participants using the decision rules described in Section 1.Footnote ⁵⁴

Because our dataset consists of cases in which certiorari was granted, there may be concerns that analysis of amicus briefs at the certiorari stage for granted cases may not be representative of all amicus briefs filed at this stage. Gathering the level of information we did on all amicus briefs in cases where certiorari was denied was not feasible. Instead, to address potential concerns, we asked the research assistants who hand-coded petitioners for Section 2 to also collect information about whether the petitions they were investigating had amicus briefs filed supporting or opposing certiorari. They did this for 4,466 randomly sampled petitions for which certiorari was denied. Only 134 of these cases (0.03%) had any amicus briefs submitted. In contrast, 254 of the 396 cases (64%) in our dataset (all cases in which cert was granted) had briefs filed at the certiorari stage. This is consistent with previous studies finding that amicus briefs at the certiorari stage are highly correlated with grants of certiorari (Reference Caldeira and WrightCaldeira & Wright, 1988; Reference Schoenherr and BlackSchoenherr & Black, 2019) and suggests that our dataset contains the majority of cert-stage amicus briefs filed. However, we wanted to be sure that there were not systematic differences in who was participating in amicus briefs at the cert stage in those cases that were granted and those that were not, so we gathered amicus data on the 134 cases with cert-stage briefs from the un-granted random sample for comparison. Generally, there are fewer briefs per case in the un-granted sample – the mean number of briefs for these 134 cases was just 2.4 briefs, and the median was 1. Generally the patterns with respect to who is participating on these briefs look similar to our analysis below, though the number of participants per case in those where cert was granted are generally higher.Footnote ⁵⁵

3.3 Analysis

Amicus briefs are crafted in response to cases already being considered by the court, either at the certiorari or merits stage. As a result, raw numbers of amicus briefs filed are not as useful as the raw numbers of petitions for certiorari used in Section 2. Instead, because the number of cases the Court agrees to hear varies from year to year, the most informative measure is the number of briefs filed per case. Standardizing by case allows us to make an apples-to-apples comparison across our different periods of interest. For this reason, the analyses that follow use the average number of amicus participants of each type per case in each period as the basis for comparison.

3.4 Political Disadvantage Analysis

As with petitioners, we expect that the most politically disadvantaged groups – those with negative social constructions and low power – will participate less in amicus filings as each Trump appointee is confirmed and the Court becomes more conservative, particularly at the certiorari stage when the Court has not yet decided whether to take the case. Figure 10 displays the average participation by group type at both the certiorari and merits stages.

Figure 10 Filings by political disadvantage pre- and post-candidate confirmation, by stage

Amicus brief participation is much lower overall at the certiorari stage than at the merits stage, and patterns do appear to differ by group. Participation by groups with low power only is practically nonexistent at the cert stage. Groups who are both low power and negatively constructed start low and show a decline as Trump’s nominees are confirmed, as Hypothesis 1 predicted.Footnote ⁵⁶ Meanwhile, the not disadvantaged show a steady increase in amicus participation at this stage. Participation in amicus briefs at the merits stage follows a slightly different pattern. Like at the cert stage, the most advantaged groups increase participation in merits briefs with every Trump appointment. Groups with either negative construction or low power generally increase, but the pattern is not as steady – there are occasional dips at the merits stage. In contrast, the most disadvantaged groups follow a weak bell curve. They increase very slightly in merits participation after Justices Gorsuch and Kavanaugh’s confirmations, then fall after Justice Barrett’s confirmation. These patterns are mostly consistent with Hypothesis 1, and also support the intuition that there are slightly different strategic considerations at the merits stage.

Though Figure 10 demonstrates some differences between groups with different levels of political disadvantage, these differences may not be solely a result of amicus strategy. It is possible that the changing composition of the petitioners for certiorari described in Section 2 changed the opportunities for different types of groups to participate in amicus briefs at the cert stage. Similarly, it is possible that changes in the membership of the Court led to changes in the types of cases the Court accepted, which in turn presented fewer opportunities for disadvantaged groups to participate as amici. To test for this possibility, we estimate an ordinary least squares (OLS) regression that includes issue area fixed effects drawn from the Supreme Court database (Reference Spaeth, Epstein and MartinSpaeth et al., 2022).Footnote ⁵⁷ The dependent variables are the logged number of brief participants per case in each of the four categories described above. We regress the number of brief participants on a case on three dummy variables as specified in the previous sections: Post-Gorsuch, Post-Kavanaugh, and Post-All. The excluded category is cases in our time period before any Trump appointees to the Supreme Court were confirmed. The dataset we use is a case-stage analysis: that is, for each of the political disadvantage categories, we have two regressions: one for amicus participation at the cert stage (labeled as ‘C’ in the regression table) and one for the merits stage (labeled as ‘M’). This allows us to assess differences in strategic incentives at each stage. Table 4 contains the results.Footnote ⁵⁸

Table 4 Amicus filers by political disadvantage and likelihood of filing an amicus brief at cert or merits stage

	Dependent variable:
	Logged Number Brief Participants on a Case
	Neg Con. Only (C)	Neg Con. Only (M)	Neg./Low (C)	Neg./Low (M)	Not Disadvantaged (C)	Not Disadvantaged (M)	Low Power Only (C)	Low Power Only (M)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Post-Gorsuch	$-$ 0.024	$-$ 0.107	$-$ 0.096	0.179	$-$ 0.024	0.042	0.018	0.153
	(0.031)	(0.137)	(0.107)	(0.214)	(0.042)	(0.087)	(0.216)	(0.262)
Post-Kavanaugh	$-$ 0.020	0.158	0.085	0.409 **	$-$ 0.032	0.121	0.181	0.423 *
	(0.030)	(0.129)	(0.101)	(0.203)	(0.040)	(0.083)	(0.205)	(0.248)
Post-All	0.013	0.082	$-$ 0.017	$-$ 0.042	$-$ 0.076 *	$-$ 0.008	0.452 **	0.310
	(0.032)	(0.140)	(0.109)	(0.220)	(0.044)	(0.090)	(0.222)	(0.269)
Observations	372	372	372	374	372	372	374	374
$R^2$	0.020	0.197	0.081	0.173	0.078	0.261	0.103	0.110
Adjusted $R^2$	$-$ 0.024	0.161	0.040	0.136	0.036	0.228	0.063	0.071

Note: * $p<0.1$ ; ** $p<0.05$ ; *** $p<0.01$

Issue area fixed effects included.

Once the issue areas the petitioners are filing in or the Court is hearing are accounted for, support for Hypothesis 1 disappears. The decreases in participation by amici who have negative social constructions and low power at the cert stage that we observed in Figure 10 are not statistically significant. At the merits stage, this group actually shows an increase in amicus participation after Justice Kavanaugh’s confirmation, though just in that period. Similarly, groups with low power but not negative construction increase amicus participation at the cert stage after Justice Barrett’s confirmation and at the merits stage after Kavanaugh’s. The increases in the number of amicus participants per case for those who are not disadvantaged under our coding scheme were dramatic in Figure 10, but appear to be entirely related to changes in the issues involved in the case. This group actually decreases participation at the cert stage after Barrett’s confirmation, and no perceived increase at either stage turns out to be significant in the regression models. This analysis suggests that the differences we see between groups in Figure 10 may be driven more by the strategic decisions of petitioners and justices than by those of amici.

3.5 Repeat Player Analysis

Hypothesis 2 predicted that repeat players in litigation – businesses, interest groups, and governments – would increase participation in amicus briefs at both the cert and merit stages as the Trump justices were confirmed. Hypothesis 3 predicted a decrease in individual participation at the cert stage. Figure 11 shows the changes over time for each of these groups by filing stage. At the cert stage, support for our expectations is mixed. Governments and interest groups do increase participation in amicus briefs as expected, though the increase for governments is not as steady. Contrary to expectations, businesses decrease participation in cert amicus briefs as the Court becomes more conservative and individuals actually increase participation. The increase in individual participation is dramatic; it more than quadruples from before any Trump justice is confirmed to after they are all confirmed. Note that this is one area where our comparison of the cases granted certiorari and those from our random sample of cases where cert was denied show differences. In the denied cases, the average number of business filers per case with amici was about a third higher than in the granted cases. The opposite is true for individuals – the average number of individual participants on a brief at the certiorari stage is three times larger in granted cases than in the denied cases we sampled. This suggests the findings for businesses and individuals may have been different if all amici at the cert stage were able to be included.

Figure 11 Filings by repeat player type pre- and post-candidate confirmation, by stage

At the merits stage, all types increase participation in amicus briefs after Justice Gorsuch’s confirmation, and most except businesses continued to increase after Justice Kavanaugh’s, albeit more modestly. All types except individuals decreased amicus participation after Barrett’s confirmation. This decrease was most dramatic for businesses. As we saw at the cert stage, individuals exhibit the greatest increase in amicus participation: their participation grows nearly fivefold from before all of Trump’s appointment until after all three confirmations.

Are issue areas driving these unexpected results? As we did with the political disadvantage analysis, we estimate an OLS model regressing the number of brief participants per case on binary indicators for post-Gorsuch’s confirmation, post-Kavanaugh, and post-all confirmations with issue area fixed effects, with separate models for the cert and merits stages. Table 5 shows the results. It suggests that the decline in participation by businesses at the cert stage is not being driven by the issue areas the Court is choosing to hear. Business participation decreases by large and significant levels after each Trump nominee is confirmed. This is counter to Hypothesis 2. Results for the other two repeat player categories, governments and interest groups, are more mixed. Consistent with Figure 11, government actors do increase their amicus participation at the cert stage, though only the increases after Justices Gorsuch and Barrett’s confirmations are significant. The growth in interest group participation at the cert stage is only significant after Barrett’s confirmation. Once issue area is controlled for, individuals do not appear to increase participation at the cert stage, but they also do not decrease, as predicted by Hypothesis 3.

Table 5 Repeat player amicus filer types and likelihood of filing an amicus brief at cert or merits stage

	Dependent variable:
	Logged Number Brief Participants on a Case
	Business (C)	Business (M)	Gov. (C)	Gov. (M)	Individual (C)	Individual (M)	Interest Group (C)	Interest Group (M)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Post-Gorsuch	$-$ 0.229 ***	0.195	0.304 **	0.349	$-$ 0.096	0.573 *	0.002	$-$ 0.033
	(0.067)	(0.172)	(0.152)	(0.239)	(0.192)	(0.296)	(0.158)	(0.206)
Post-Kavanaugh	$-$ 0.191 ***	0.278 *	0.216	0.369	0.120	0.959 ***	0.203	0.263
	(0.063)	(0.163)	(0.144)	(0.227)	(0.181)	(0.280)	(0.149)	(0.195)
Post-All	$-$ 0.169 **	$-$ 0.016	0.347 **	0.410 *	0.195	0.670 **	0.271 *	0.048
	(0.068)	(0.177)	(0.156)	(0.246)	(0.197)	(0.304)	(0.162)	(0.211)
Observations	372	374	374	374	373	374	373	374
$R^2$	0.131	0.114	0.061	0.067	0.048	0.104	0.117	0.165
Adjusted $R^2$	0.092	0.074	0.018	0.025	0.005	0.064	0.078	0.127

Note: * $p<0.1$ ; ** $p<0.05$ ; *** $p<0.01$

Issue area fixed effects included.

At the merits stage, things look different. We do see a significant increase in business participation in amicus briefs, but only after Kavanaugh’s confirmation. Government participation seems to increase, but is only significant after Barrett’s confirmation. No increases in participation are evident for interest groups. Individuals, on the other hand, show significant increases after every Trump justice is confirmed.

3.5.2 Who Are the Government Filers?

To get a more nuanced look at government participation in amicus briefs, we subset the data to only government actors. Figure 12 shows the differences in government type pre- and post-Trump’s appointees’ confirmations in each stage for five types of government actors: federal, international, local, state, or tribal. International and local governments are largely absent at the cert stage, and tribal governments only became noticeably involved after Justice Barrett’s confirmation. As with petitioners, the real growth in government participation in amicus briefs at the cert stage is coming from states. It spikes after Gorsuch’s confirmation before leveling off at over twice its initial rate. At the merits stage, all types of government are more engaged, with state and local governments as the most active participants. Local and state governments more or less double their participation in briefs filed per case after Gorsuch’s confirmation and beyond.

Figure 12 Filings by government actors pre- and post-candidate confirmation, by stage

The federal government appears to have low but relatively steady participation at both the cert and merits stages, but this low level of participation is misleading for two reasons. First, there is only one federal government, and it is usually represented by the Office of the Solicitor General (OSG). Even though there are many federal government agencies, the OSG files amicus briefs only as the United States. Therefore, there is a ceiling for the number of amicus participants per case in the federal government category that is much lower than for other government types. Second, the federal government is much more often a party to cases before the Court than other government types. During our period of study, the federal government is a party – either the petitioner or the respondent – in 139 out of the 396 cases the Court heard. Litigants do not act as both parties and amici in the same case. When considered in this context, the federal government’s participation before the Court is remarkably high, consistent with its repeat player status. It participates as a party in 35% of the cases the Court hears and as an amicus in approximately 40% of the cases, meaning that it is involved in approximately 75% of the cases considered by the Supreme Court during our period of study.

Table 6 shows the OLS estimations of these relationships with issue area fixed effects. Consistent with Figure 12, at the certiorari stage, the only type of government that consistently shows significant growth is state governments, and their increase in participation is significant after each Trump justice’s confirmation. Tribal governments’ amicus filings after Barrett’s confirmation are also significant. At the merits stage, the only statistically significant increases are for local governments after Barrett’s confirmation, and states after Kavanaugh’s. Thus, it appears that state government actors are the ones changing their behavior most in response to the Court’s changing membership, but which states? We address that next as we test our ideological hypotheses.

Table 6 Government amicus filers and likelihood of filing an amicus brief at cert or merits stage

	Dependent variable:
	Logged Number Brief Participants on a Case
	Federal (C)	Federal (M)	Inter-national (C)	Inter-national (M)	Local (C)	Local (M)	State (C)	State (M)	Tribal (C)	Tribal (M)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
Post-Gorsuch	$-$ 0.002	$-$ 0.066	0.005	0.058 *	0.001	0.164	0.302 **	0.345	0.016	0.064
	(0.034)	(0.055)	(0.014)	(0.032)	(0.014)	(0.134)	(0.151)	(0.237)	(0.027)	(0.078)
Post-Kavanaugh	$-$ 0.037	0.017	$-$ 0.005	$-$ 0.008	0.021	0.164	0.255 *	0.406 *	$-$ 0.002	$-$ 0.002
	(0.032)	(0.052)	(0.013)	(0.031)	(0.014)	(0.127)	(0.143)	(0.225)	(0.026)	(0.074)
Post-All	$-$ 0.045	$-$ 0.032	$-$ 0.010	$-$ 0.014	0.010	0.244 *	0.352 **	0.401	0.059 **	0.081
	(0.035)	(0.056)	(0.014)	(0.033)	(0.015)	(0.138)	(0.155)	(0.244)	(0.028)	(0.080)
Observations	372	373	372	373	372	372	374	374	372	372
$R^2$	0.169	0.084	0.017	0.051	0.049	0.089	0.053	0.073	0.052	0.067
Adjusted $R^2$	0.132	0.043	$-$ 0.027	0.008	0.007	0.048	0.010	0.031	0.010	0.025

Note: * $p<0.1$ ; ** $p<0.05$ ; *** $p<0.01$

Issue area fixed effects included.

3.6 Ideology Analysis

We subset the data even further to only state government actors to examine the partisanship of these state governments. Recall that we code state government actors into three categories: blue, red, or purple, based on whether the attorney general and governor are both Democratic (blue), both Republican (red), or belong to different parties (purple). Figure 13 shows the distribution of state government amicus participants across these categories in each period by stage.

Figure 13 Filings by state government actors pre- and post-candidate confirmation, by stage

At the cert stage, blue and purple states look quite similar. They participate at low rates before any Trump nominee is confirmed, then their participation spikes – by 300% and 400%, respectively – following Gorsuch’s confirmation, perhaps reflecting a reaction to early policies of the Trump administration. Both drop off after Kavanaugh’s confirmation, with blue states reverting to about where they started and purple states slightly higher. The pattern for red states at the certiorari stage shows a similarly dramatic jump after Gorsuch’s appointment, but then stays at four times the initial level. Comparisons between cert stage petitions in denied and granted cases in Appendix Section C suggest that red states’ increased amicus participation may actually be understated here. The average denied case had three more red state participants at the cert stage than the average granted case. In contrast, blue states participated at a much lower rate in cases that were denied.

At the merits stage, the patterns are more complicated. Blue states increase their amicus participation significantly after the first two confirmations of Trump justices and drop slightly after Barrett’s confirmation. The increase in blue state participation is the largest of all state types, more than doubling between the beginning and end of our period of study. Red states have a higher level of participation at the merits stage overall, but the growth in their participation is not proportionally as large, and it drops back to its initial level after Barrett’s confirmation. The pattern for purple states at the merits stage looks similar to its pattern at the cert stage – a big spike in participation after the Gorsuch confirmation, and then a decline.

To understand whether and to what extent these patterns are suggestive of strategic behavior, we need to put them in context with the changing partisanship of the states themselves. At the beginning of our period of study (the pre-Gorsuch period), there were about twice as many red states in the country as either blue or purple states by our coding. In this period, amicus participation by states at both stages is therefore roughly proportional to the partisan make-up of the states. However, over time, several red and purple states flipped to blue, and comparatively few went the other way. By the end of our period of study, there were only 8% more red states than blue states. Yet in that same period of time, amicus briefs by red states at the certiorari stage quadrupled, while those from blue states spiked briefly after Gorsuch then returned to their pre-Trump levels. This suggests that conservative states were more likely to perceive the Supreme Court as open to their interests, and sought to get those interests on the Court’s agenda. In contrast, at the merits stage, we see blue states’ participation growing more rapidly than red states. This may in part be a result of the growing number of blue states over time, or it may reflect a strategic effort to counteract the red states’ influence.

Table 7 shows the regression results of the subset of state governments by partisanship with issue area fixed effects. Controlling for issue area, we see large and statistically significant increases in red state filings after every Trump justice confirmation at the cert stage and after Gorsuch and Kavanaugh’s confirmations at the merits stage. Blue states show neither growth nor decline at the cert stage but show even greater growth than red states at the merits stage after the confirmations of both Kavanaugh and Barrett. Purple states show modest growth in participation at the cert and merits stages, but only after Gorsuch’s confirmation.

Table 7 State government actor amicus filers and likelihood of filing an amicus brief at cert or merits stage

	Dependent variable:
	Logged Number Brief Participants on a Case
	Blue (C)	Blue (M)	Purple (C)	Purple (M)	Red (C)	Red (M)
	(1)	(2)	(3)	(4)	(5)	(6)
Post-Gorsuch	0.061	0.151	0.176 **	0.219 *	0.266 **	0.360 *
	(0.066)	(0.164)	(0.074)	(0.129)	(0.127)	(0.186)
Post-Kavanaugh	0.020	0.423 ***	0.106	0.127	0.256 **	0.315 *
	(0.062)	(0.155)	(0.071)	(0.122)	(0.120)	(0.176)
Post-All	$-$ 0.002	0.312 *	0.088	$-$ 0.039	0.229 *	0.026
	(0.067)	(0.169)	(0.077)	(0.132)	(0.130)	(0.191)
Observations	373	374	374	374	374	374
$R^2$	0.050	0.093	0.052	0.077	0.054	0.074
Adjusted $R^2$	0.008	0.052	0.010	0.035	0.012	0.033

Note: * $p<0.1$ ; ** $p<0.05$ ; *** $p<0.01$

Issue area fixed effects included.

3.6.1 Interest Group Ideology

Finally, we move from considering state partisanship to examination of our hypotheses regarding interest group ideology. We subset to only interest groups and code the approximately 6,000 groups into three mutually exclusive categories: liberal, conservative, or nonpartisan. Recall that liberal groups are those that explicitly mention a liberal or progressive agenda on their websites or other public material and conservative groups mention conservative or family values. All those not coded as explicitly conservative or liberal were labeled as nonpartisan. Most groups are not explicitly partisan.

Figure 14 shows the distribution of briefs filed per case by interest group ideology. Both conservative and nonpartisan interest groups increased their rate of amicus participation over our time frame at the cert stage, but there was especially great growth among conservative interest groups. We see very little cert stage participation in amicus briefs by explicitly conservative groups before Justice Gorsuch’s confirmation. Their participation doubles after Gorsuch is confirmed, nearly doubles again after Kavanaugh’s confirmation, and after Barrett’s confirmation, they are participating at ten times their pre-Gorsuch levels. In contrast, there is no noticeable participation by liberal interest groups at all at the cert stage. Nonpartisan interest group participation grows steadily, but not exponentially like that of the conservative groups.

Figure 14 Filings by interest group type pre- and post-candidate confirmation, by stage

Explicitly liberal groups are more active at the merits stage, but are still overshadowed by their conservative counterparts. In the pre-Gorsuch period, participation by liberal and conservative groups is fairly similar – liberal groups file an average of 0.6 briefs per case, and conservative groups file an average of 0.8. However, after Gorsuch’s confirmation, conservative groups’ participation more than triples, and after Barrett is confirmed, it is more than four times its pre-Gorsuch rate. Participation by liberal groups does show an increase, but it is not nearly as large or dramatic. The patterns for liberal groups are more similar to those for nonpartisan groups, which approximately double their participation then fall off somewhat after Barrett’s confirmation.

Table 8 shows the regression results among interest group filers with issue area fixed effects. We see strong support for Hypothesis 4; conservative interest groups are increasing their amicus participation at both the cert and merits stages as each Trump nominee is confirmed, especially once Kavanaugh and Barrett reach the court. While we do not see direct support of Hypothesis 5 because liberal groups are not decreasing in their overall rate of amicus participation, their participation is decreasing as a proportion of all amicus filers over this period. Given that ideological representation was virtually balanced previously (Reference Abi-Hassan, Box-Steffensmeier, Christenson, Kaufman and LibgoberAbi-Hassan et al., 2023), this rapid growth in participation by conservative groups is striking.

Table 8 Interest group amicus filers and likelihood of filing an amicus brief at cert or merits stage

	Dependent variable:
	Logged Number Brief Participants on a Case
	Conservative (C)	Conservative (M)	Liberal (C)	Liberal (M)	Nonpartisan (C)	Nonpartisan (M)
	(1)	(2)	(3)	(4)	(5)	(6)
Post-Gorsuch	0.052	0.288 **	0.006	$-$ 0.016	$-$ 0.042	$-$ 0.048
	(0.084)	(0.129)	(0.019)	(0.097)	(0.147)	(0.199)
Post-Kavanaugh	0.162 **	0.359 ***	$-$ 0.012	0.137	0.137	0.235
	(0.079)	(0.122)	(0.018)	(0.091)	(0.139)	(0.189)
Post-All	0.178 **	0.269 **	$-$ 0.002	0.111	0.214	$-$ 0.024
	(0.086)	(0.133)	(0.020)	(0.099)	(0.151)	(0.205)
Observations	373	374	372	373	373	374
$R^2$	0.133	0.230	0.042	0.177	0.105	0.156
Adjusted $R^2$	0.094	0.196	$-$ 0.001	0.140	0.065	0.118

Note: * $p<0.1$ ; ** $p<0.05$ ; *** $p<0.01$

Issue area fixed effects included.

3.7 Summary and Conclusions

The number of amicus participants grew dramatically during our period of study, from about 3,000 participants in 2016 to over 14,000 in 2021, while the number of briefs grew slightly. Growth in amicus participation was not consistent across different types of groups, however. Our most notable finding is that, consistent with our expectations, amicus participation by conservative interest groups and red state governments increased dramatically. This suggests that those with explicitly conservative policy preferences expected to find a more receptive audience as the Court’s conservative majority solidified with each Trump appointee’s confirmation.

The results for participation by other groups are more nuanced. Our analysis of the most politically disadvantaged groups shows an apparent decrease in amicus participation, but when we control for issue area, this decrease is not significant. This suggests the decline in amicus participation by the most politically disadvantaged groups is being driven more by the types of cases petitioners are filing and the Court is granting rather than by the strategic decisions by these groups are making as amici. Repeat player theory proves to be a relatively poor fit for predicting amicus behavior. Perhaps this is because amicus participation is dominated by repeat players, and most individuals who participate in amicus briefs are policy-interested elites.

Our expectations regarding differing strategic incentives at the certiorari and merits stage are largely borne out. We see more and stronger evidence of strategic behavior at the certiorari stage. Notably, no group type that we examined exhibited a decline in amicus participation at the merits stage. However, participation by politically disadvantaged groups, liberal interest groups, and blue states grew less dramatically than participation by other groups. These differences matter. Justices are known to sometimes pull arguments and even direct quotes from amicus briefs. Conservative interest groups and states are rushing forward to provide arguments the Court can use and may not be effectively countered by more liberal or moderate perspectives.

4.1 Beyond Trump’s Justices

We have demonstrated not just that appointees matter, but that both the individual and cumulative effects of new justices are consequential for litigant and amicus behavior. Cumulative effects raise the stakes for elections and subsequent judicial nominations. Although our analysis focuses on Supreme Court appointments by Donald Trump, it has broader significance. It demonstrates that any president who gets enough nominations can shift the Court and the opportunities for different types of litigants dramatically.

As of August 2022, President Joe Biden had appointed more federal judges than any president since John F. Kennedy over the same amount of time into an administration.Footnote ⁶³ Biden’s differences from his predecessors are not just notable in terms of the number of appointments. His appointees are also more diverse than those of any other president in history, in terms of race, sex, sexual orientation, background, and other metrics (Reference Fredrickson and NeffFredrickson & Neff, 2022). His Supreme Court pick, Ketanji Brown Jackson, is the first Black woman on the highest court, the third Black person, the sixth woman, and the first former public defender. This choice has significant implications not just for public perceptions of the Court, but also, as we have shown, for litigant and amicus strategy. Following our findings, it is possible that there will be additional changes in litigant and amicus behavior following Justice Jackson’s confirmation. However, the overwhelming size and conservatism of the current majority make it unlikely that the appointment of an individual justice in the liberal minority coalition will shift things much. Instead, additional appointments will likely be necessary to reverse current trends given today’s conservative supermajority.

The prospects for change look somewhat different when we consider the new, large slate of nominees and confirmations to the lower federal courts. We suspect shifts in strategic incentives similar to those we have documented at the Supreme Court level may be occurring as Biden’s picks take the bench in the lower courts, particularly the US Courts of Appeals. This is an important avenue for future research. There are limits to how far lower court judges can stray from Supreme Court preferences in salient issue areas, but given the tiny proportion of cases the Supreme Court hears each year, there is at least some opportunity for lower courts to make room for different participants and their interests.

4.2 Future Directions

Another fruitful area for future research is further inquiry into the precise mechanisms that shape litigant and amicus incentives. The strength of our findings regarding the increased mobilization of conservative states and interest groups as litigants and amicus participants suggests that expectations regarding the justices’ policy preferences are a primary motivator, but they are likely not the only influence. For example, it does not explain our peculiar findings regarding a decrease in petitions for certiorari and in amicus activity by businesses. Conservatives have traditionally been thought to favor pro-business policy, so why were businesses turning away from the Court during this period? Did businesses find these particular conservative justices unwelcoming to their interests? Did greater opportunities to influence policy in other branches of government develop and lead businesses to shift their energies and resources in those directions? Or did businesses simply assume that with a Court so closely aligned with their preferences and interest groups actively making their case, they did not have to invest time and money to get the outcomes they want?

Additionally, are there other characteristics beyond ideology that might also shape participation before the Court. For example, looking at Biden’s appointees, could a judge’s or justice’s background, race, or age matter? Existing research shows that judges’ individual characteristics may lead to differences in support for the courts (Reference Achury, Casellas, Hofer and WardAchury et al., 2023; Reference Badas and StaufferBadas & Stauffer, 2018; Reference Krewson and OwensKrewson & Owens, 2021), but more work is needed on how these changes in public opinion translate into differences in litigant behavior. That is, if a potential litigant is supportive of a co-racial nominee (Reference Kaslovsky, Rogowski and StoneKaslovsky, Rogowski, & Stone, 2021), for example, might that mean they then use that information to also decide whether to file a case, appeal it to the Supreme Court, or become involved in an amicus brief? Would a Black woman who sees Justice Jackson on the Supreme Court bench be more likely to move forward with a legal claim, all else equal? While we have shown differences in participatory behavior, future research ought to investigate the range of mechanisms or reasons litigants or amicus participants have for altering their behavior in the courts.

These precise mechanisms may also differ across levels of the federal hierarchy. While we focus on the Supreme Court in this Element for several reasons – data availability, the publicized nature of Supreme Court nominations, and the wide-ranging impact of Supreme Court decisions – district and circuit courts may also be avenues for changing litigant behavior. Namely, we would expect savvy litigants to take note of current preferences at their local and regional district and circuit courts in addition to the Supreme Court. Future work ought to explore these differences further to illuminate patterns in the oft-understudied lower courts (Reference BoydBoyd, 2016; Reference Hübert and CopusHübert & Copus, 2022; Reference Martinek, Kemper and Van WinkleMartinek, Kemper, & Van Winkle, 2002; Reference Massie, Hansford and SongerMassie, Hansford, & Songer, 2004).

We also invite consideration of a normative implication of this research. Ought we be concerned if, with each new justice confirmed to the court, we see significant differences in litigant and amicus behavior? Should there be relatively equal levels of participation and perceived access to the courts over time? Is it problematic that one new justice (or, in our case, three new justices) can have such a significant influence on who comes to the court? Our research dovetails nicely with the growing literature on the politicization of confirmation hearings and judicial selection (Reference Armaly and LaneArmaly & Lane, 2023; Reference BadasBadas, 2023; Reference Collins Jr, Ringhand and BoydCollins et al., 2023; Reference Rogowski and StoneRogowski & Stone, 2021) and encourages scholars not only to consider how these political battles influence public opinion of the courts, but also how they might affect representation and policy outcomes downstream from successful confirmations.

4.3 Final Thoughts

This project brings together several different strands of literature on law and courts and demonstrates how they can and should speak to one another. Early on, political disadvantage theory posited that the courts were a particularly fruitful policymaking venue for those lacking traditional political power (Reference CortnerCortner, 1968; Reference VoseVose, 1959). Later studies suggested that businesses and governments may expect even greater benefits from litigation than disadvantaged groups (Reference GalanterGalanter, 1974; Reference OlsonOlson, 1990). All of these studies assume that litigants are behaving strategically, but seem to hold the Court constant. Existing explanations do not take seriously the make-up of who is on the Court and how that might shape strategic incentives. Yet, we know from studies of judicial decision-making that judges have their own policy priorities and strategic considerations (Reference Epstein and JacobiEpstein & Jacobi, 2010; Reference Segal and SpaethSegal & Spaeth, 2002). If litigants are making strategic calculations, they should consider judicial preferences as part of that analysis. For example, when the Court exhibits pro-business policy bias (Reference WhitehouseWhitehouse, 2015), it should make businesses more likely to view the Court as a good bet for policymaking, and should raise caution among advocates for regulation. Studies of amicus curiae behavior at the Supreme Court do a better job of considering the make-up of the Court (Reference Bils, Rothenberg and SmithBils et al., 2020; Reference HansfordHansford, 2004a), but these studies focus more on the influence of the amici on the Court rather than the influence of the Court’s membership on amicus behavior (Reference CollinsCollins, 2004; Reference Mann and FronkMann & Fronk, 2021; Reference McCammon, Sudibjo, Beeson-Lynch, Brockman and MoonMcCammon et al., 2022; Reference Songer and SheehanSonger & Sheehan, 1993). We bring these strands together by considering how both litigants and amicus participants respond to dramatic changes in the make-up of the Supreme Court.

Shifts in litigant behavior matter because they result in shifts in policy outputs. Courts have often played a minoritarian function, and reduction or withdrawal of cases involving the interests of disadvantaged groups can increase political inequality (Reference Gibson and NelsonGibson & Nelson, 2021). Older studies have documented the link between litigant behavior and policy change with respect to particular policy areas (see, e.g., Reference KobylkaKobylka, 1987). More recent literature has done the same with amicus behavior (Reference McCammon, Sudibjo, Beeson-Lynch, Brockman and MoonMcCammon et al., 2022). Our work builds on these studies by taking a more comprehensive view, looking at both litigant and amicus curiae across all policy areas. We also contribute to the literature on group representation in the policymaking process more broadly. While most groups use a wide range of advocacy activities, litigation is one of the least frequently used tactics (Reference GrossmannGrossmann, 2012; Reference Nownes and FreemanNownes & Freeman, 1998). Understanding how changes in the make-up of the Court affects access to this important policymaking venue advances our overall understanding of who gets what and when and how (Reference LasswellLasswell, 1958). The voices present at the Court can significantly shape policy outcomes and if certain voices are absent, that has important consequences for whether and to what degree policy is representative of the concerns and preferences of an increasingly diverse nation.