TRENDS: The Influence of Personalized Knowledge at the Supreme Court: How (Some) Former Law Clerks Have the Inside Track

Abstract

When arguing at the U.S. Supreme Court, former High Court law clerks enjoy significant influence over their former justices. Our analysis of forty years of judicial votes reveals that an attorney who formerly clerked for a justice is 16 percent more likely to capture that justice’s vote than an otherwise identical attorney who never clerked. What is more, an attorney who formerly clerked for a justice is 14 to 16 percent more likely to capture that justice’s vote than an otherwise identical attorney who previously clerked for a different justice. Former clerk influence is substantial, targeted, and appears to come from clerks’ personalized information about their justices. These results answer an important empirical question about the role of attorneys while raising normative concerns over fairness in litigation.

Keywords

Supreme Court judicial politics attorneys law clerks lobbying matching

Supreme Court law clerks hit the jackpot—in more ways than one—when they departed the Court at the close of the 2018 term. Most of them entered private practice, where they collected signing bonuses of $400,000 (Mauro 2018). These bonuses, combined with their hefty salaries, allowed former clerks to earn upward of three-quarters of a million dollars in their first year away from the Court (Ward, Dwyer, and Gill 2014). Law firms throw money at former clerks for the same reason companies, unions, and organized interests hire former government officials as lobbyists: they expect these insiders to influence their previous employers. “High powered firms perceive that the clerk experience can give their clients a distinct advantage before the nation’s highest court . . .” (O’Connor and Hermann 1995, 247). As one partner at a large firm commented, “We have found [that hiring former clerks is] a terrific investment for us . . . They have this unique view of how judges think, of how the justices interact” (Kendall 2012, 2). Or so the story goes.

Do former clerks actually influence the justices? If so, why? Despite a marked increase in former clerks who enter private practice—and Supreme Court practice in particular—scholarship is surprisingly silent on whether and why such influence exists. In what follows, we draw from literature on revolving door lobbying and employ matching analyses to examine whether and why former Supreme Court law clerks win the justices’ votes.

Our analysis shows that former clerks provide a significant but highly targeted advantage in terms of capturing justices’ votes. More specifically, an attorney who formerly clerked for a justice is 16 percent more likely to capture that individual justice’s vote than an otherwise identical attorney who never clerked. This former clerk, however, has no significant advantage when trying to sway the votes of other justices. What this means is that former clerk influence is real, but targeted, and that former clerk success largely derives from the personalized information they know about their justices.

These results have a number of important implications. Former clerks regularly litigate at the Court. Since the start of the Roberts Court era, fully 75 percent of all argued cases have observed at least one former law clerk appear. Their role, in other words, is pervasive. Moreover, the fact that inside information leads former law clerks to capture a justice’s vote seems normatively problematic. Greater legal acumen and stronger legal arguments should lead to success. The ability to “see the man behind the curtain” through personal exposure to decision makers should not. More broadly, the results also offer insight into revolving door lobbying success. Lobbying surely is not the same as litigation; nevertheless, the logic of our findings suggests that revolving door lobbyists, like former clerks, may succeed because they capitalize on personalized information that few others enjoy.

The Importance of Lawyers before the U.S. Supreme Court

Scholarship shows that attorneys play a crucial role throughout the Supreme Court’s decision-making process. Typically, findings demonstrate that increasingly capable and experienced attorneys are more likely to win. For example, McGuire (1995) argues that attorneys with greater experience are more likely to win their cases. As he put it, “the litigation experiences of the counsel who represent [parties] are a significant determinant of judicial outcomes” (McGuire 1995, 187). McGuire’s (1993a) earlier research on Washington elite lawyers tells a related story. Similarly, Szmer and Ginn (2014) find that more capable attorneys are more likely to influence justices in cases where the justices have less expertise (see also Szmer, Johnson, and Sarver 2007). Looking at asylum cases, Miller, Keith, and Holmes (2015) find that attorneys with past success generally are more likely to win in the future.

Looking more specifically at the brief writing stage, Black et al. (2016) find that justices are more likely to vote for credible attorneys—those whose briefs eschew emotionally charged language. Long and Christensen (2008) explore how the use of intensifiers (i.e., words such as clearly, obviously, and very) in parties’ briefs correlates with their chances of success. The more intensifiers the brief employs, the less credible the attorney appears and, in turn, the less likely the attorney is to succeed. Corley (2008) finds that justices borrow more language from the Solicitor General’s briefs, from Washington elite attorneys, and attorneys with more experience—that is, the most credible and capable of all attorneys (see also Wedeking 2010).

Looking at oral arguments Johnson, Wahlbeck, and Spriggs (2006) analyze the private records of Justice Blackmun to determine whether oral argument influences case outcomes. Using Blackmun’s private ratings of the attorneys to measure argument quality, Johnson and colleagues find a correlation between better quality oral arguments and the likelihood of victory for that party. Attorneys whom Blackmun rated as having done a better job were more likely to win their case—and this result held even while controlling for instances in which Blackmun might simply be inclined ideologically to award the attorney a good grade.¹

Taken together, these studies suggest that attorney influence comes from the ability to provide justices with information they know justices need. Sometimes this information is legal, with attorneys providing useful insight into precedent and statutory language (Schoenherr and Black 2019; Black and Owens 2009). Other times, the attorneys provide policy-based information relating to the ideological dimensions of a case (Bailey, Kamoie, and Maltzman 2005). But the bottom line is that attorneys who can provide justices with useful and trustworthy information tend to perform well.

It is surprising, therefore, that few studies specifically examine the influence of former Supreme Court clerks. After all, clerks were at one point among the justices’ most trusted staff. Furthermore, having been hired at one of the most elite institutions in the world, they would appear to have the skills to be excellent attorneys. And given their time working so closely with justices, they surely have information that could provide them insight into their justices’ thoughts. If attorneys can influence the Court, it would seem that attorneys who once worked for the Court would have a leg up.

Yet, only a handful of studies examine former clerks. Some point out that former clerks are likely to return to argue before the Court (McGuire 1993a, 1993b; O’Connor and Hermann 1995), but only a small number examine whether they are more likely to succeed. And even those studies generate mixed results (e.g., Johnson, Wahlbeck, and Spriggs 2006; McGuire 2000). Moreover, it is unclear just what causes former clerks to win. Is it skill? Resources? Personalized knowledge? All of the above? We seek to answer these questions. To inform our understanding of precisely why former law clerks may influence justices, we look to scholarship on revolving door lobbying.

Revolving Door Lobbying and Theories of Former Law Clerk Influence

Revolving door lobbying is “the process by which [governmental workers] seamlessly leave government service to represent private clients before the very same congressional [or other] offices for which they previously worked” (Cain and Drutman 2014, 27).² Put simply, the concept focuses on people who once were “on the inside” of government but who are now “on the outside” seeking to influence policy makers. Scholars have discovered that revolving door lobbying is highly successful, and they have proffered a number of theories to explain this success. We summarize each and offer an application to the study of former law clerks.

Substantive Expertise

This theory focuses on domain-specific content knowledge individuals acquire from working on policy issues (Esterling 2004). Revolving door lobbyists are successful because they have become experts about the subject at hand and can convey their knowledge to policy makers. While working in government, they tend to accrue substantive expertise they later deliver to policy makers. Staffers whose members serve on committees quickly become experts on those issues. Later, when they return to lobby, they use that information to their advantage. And policy makers come to trust the information they receive from them. Indeed, because members of Congress tend to be policy generalists, they need assistance from policy experts (Esterling 2004) and use the information they receive from them.

Applied to former clerks, the logic is fairly straightforward. To assist their justices, law clerks must learn the nuances of every legal issue that comes before the Court. As Peppers (2006, 145) puts it, the modern law clerk “is relied on to master complex areas of the law . . .” Clerks are “trusted decision makers who work alongside the nation’s highest officials” (Ward and Weiden 2006, 24). Given the breadth of issues they must learn, and their importance to justices, the position is “more responsibility than [they’ll] ever have again” (Ward and Weiden 2006, 3). Peppers (2006, 205) states that “at no other time in the history of the Supreme Court have law clerks been given as many substantive responsibilities” than they have on the modern Court. It could be, then, that former clerks win so many cases because their clerkship duties required them to become experts on the law.

Wealth and Resources

The wealth and resources theory suggests that revolving door lobbyists are highly successful because they enjoy massive resources that they employ to support their lobbying efforts. When they leave government service, they often work for large lobbying firms with the resources to amass strong arguments. So when these revolving door lobbyists return to lobby their former colleagues, they enjoy considerable resources—and it is those resources that line up to influence policy makers, not some halo of knowledge. In this vein, Baumgartner et al. (2009) find that businesses are slightly more likely to accomplish their lobbying goals than non-businesses. Their resources allow them to make persuasive arguments, which, in turn, become policy.

This theory would suggest that former clerks tend to win more often than non-clerks because they get hired at the best law firms in the country and enjoy the resources to compile good trial court records, hire excellent experts, and sink the resources into mastering every legal issue in their cases. Former High Court law clerks would work for clients who pay whatever it takes to win and for firms that have the resources to do the same.

Process Expertise

According to this theory, the revolving door lobbyist succeeds so often because she knows how to work the gears of the policy-making institution (Drutman 2014). The lobbyist knows how the decision-making process operates from start to finish and where lobbying efforts are most likely to be efficacious. The process expert is not successful because she knows policy inside and out (though she may); she knows which member of Congress to lobby, which regulator to contact, or which staffer actually gets things done. As LaPira and Thomas (2017, 6) note, “revolving door lobbyists’ insider knowledge about the decision making procedures and processes of their former employers can minimize the numerous political risks borne by their clients.” One former member of Congress stated, “I was a . . . legislator for twenty-four years. I know how that process works” (LaPira and Thomas 2017, 36).

A similar logic applies to former High Court clerks. As one lawyer stated, “requiring our new lawyers to have completed a federal clerkship assures that they have unique insight into how judges work . . .” (Lat 2015). Former clerks would know, for example, how the Court goes about setting its agenda as well as the kinds of things over which justices bargain when crafting opinions. Former clerks know what goes on behind the scenes at the Court. And they leverage that knowledge about the process to make well-crafted arguments.

Personalized Information

Finally, the personalized information theory holds that the revolving door lobbyist is so successful because she knows what matters to an individual policy maker. A member of congress may be interested in how one policy will influence certain manufacturers in her district. Other policy makers may have pet peeves that only insiders know. One study put it thusly:

[Revolving door lobbyists] know what makes different key people tick—what they are likely to do in response to what, and how best to approach them . . . having these relationships also makes former staffers privy to intelligence that others may lack. (Drutman 2014, 89)

They can use all of these bits of “relational capital” (Vidal, Draca, and Fons-Rosen 2012, 3732) when lobbying their former colleagues and employers.

Former clerks enjoy personalized information about the justices. Clerks generally work for a justice for one intense year (sometimes two; Peppers 2006). They assist their justices with every facet of the Court’s work. They help justices decide which cases the Court will review, they prepare justices for oral argument, and they write opinion drafts. As Ward and Weiden (2006, 2) put it, “[i]nformal, life-long relationships develop right away between justice and clerk” with clerks becoming a part of their justices’ extended family. A law clerk comes to know his or her justice very well. And that information can later be leveraged to create an advantage.³

Data and Measures

To examine former clerk influence, we started by using the Supreme Court Database to identify all orally argued cases (including signed opinions, per curiam opinions, and judgments of the Court) decided between the Supreme Court’s 1979 and 2018 terms (N = 3,589).⁴ We then retained only cases where two attorneys appeared at oral argument, which obtained for about 62 percent of the cases in our data. We opt for this approach because it provides the cleanest way of assessing the specific influence of an individual attorney.

The theories of revolving door influence suggest that a former clerk’s influence might vary with regard to the individual justice she is trying to persuade. For example, the implication of the process expertise theory is that all former clerks will outperform non-clerks because of their process knowledge, whereas the personalized information theory suggests that former clerks will only have a leg up in persuading their former justices. As such, our unit of analysis is the attorney-justice vote dyad per case. Our dependent variable—Vote—asks: in each case, did a justice vote for the arguing attorney under analysis in the case? If so, we code Vote as 1. If the justice voted against the attorney, we code Vote as 0.

We employ four main covariates of interest to address the attorney–justice relationship. The purpose of these variables is to tap into whether an attorney who argued a case was a former law clerk and, if so, the relationship the attorney had with the voting justice. (In a moment, we will provide an example to clarify these categories.) To determine whether an attorney was a law clerk, and for whom, we rely on Peppers (2006) and Ward and Weiden (2006). For more recent terms not included in their books, we used updated lists they provided to us, or that legal blogs—such as Above the Law—published.

Never Clerked

Our first indicator accounts for attorney–justice relationships (or, non-relationships) when the arguing attorney never clerked for the justice under analysis or for anyone else on the Supreme Court. If the attorney never clerked, we code this variable as 1 for all nine of the attorney–justice dyads in the case. Approximately 80 percent of the dyads in our data take on this value.

Clerked, Non-sitting Justice

Our next indicator captures the relationship between an attorney who once clerked and a justice who came to the Court after the clerk left. In such contexts, we code the variable for the attorney–justice dyad as 1, and 0 otherwise. This value appears in about 11 percent of the observations in the full data. Among the subset of observations when the attorney was a former clerk, it accounts for 54 percent of the observations.

Clerked, Sitting Justice

This indicator captures the relationship between an attorney who once clerked for a justice who served with the justice under analysis. In these instances, we code the variable for the attorney–justice dyad as 1, and 0 otherwise. About 8 percent of our observations take on this value in the full data. Among the subset of observations when the attorney was a former clerk, it accounts for about 40 percent of the observations.

Clerked, Employing Justice

This indicator captures instances where the attorney clerked for the justice under analysis. When this occurs, we code the variable for the attorney–justice dyad as 1, and 0 otherwise. In the full data, this value appears about 1.1 percent of the time. Among the subset of observations when the attorney was a former clerk, it accounts for about 6 percent of our observations. At first blush, this may seem to be a rare event. As we show below, however, it turns out to be quite important substantively.

To understand more clearly the differences among these variables, consider the example of Thomas G. Saunders, who clerked for Justice Ginsburg during the Court’s 2007 term. In March 2015, he argued Kimble v. Marvel Enterprises, Inc. When Saunders argued his case, he saw a mixture of familiar and unfamiliar faces. Chief Justice Roberts and Justices Scalia, Kennedy, Thomas, Ginsburg, Breyer, and Alito remained from when Saunders had clerked. But after his clerkship, Justice Sotomayor replaced Justice Souter (who served when Saunders clerked) and Justice Kagan replaced Justice Stevens (who also served when Saunders clerked).

As Table 1 shows, for both the Saunders–Sotomayor and Saunders–Kagan dyads, we assign a value of 1 to Clerked, Non-sitting Justice. For the Saunders–Roberts, Saunders–Scalia, Saunders–Kennedy, Saunders–Thomas, Saunders–Breyer, and Saunders–Alito dyads, we assign a value of 1 to Clerked, Sitting Justice. Finally, for the Saunders–Ginsburg dyad, we assign a value of 1 to Clerked, Employing Justice. The value of this coding approach is that it allows us to uncover not only whether influence exists but why.

Table 1.

Example of Attorney–Justice Coding Scheme.

Attorney–justice dyad	Clerked non-sitting justice	Clerked sitting justice	Clerked employing justice
Saunders–Roberts	0	1	0
Saunders–Scalia	0	1	0
Saunders–Kennedy	0	1	0
Saunders–Thomas	0	1	0
Saunders–Ginsburg	0	0	1
Saunders–Breyer	0	1	0
Saunders–Alito	0	1	0
Saunders–Sotomayor	1	0	0
Saunders–Kagan	1	0	0

This example portrays Thomas G. Saunders, who clerked for Justice Ginsburg during the Court’s 2007 term and later argued a case in 2015.

To ensure our results recover the effect of being a former law clerk—and not some unmeasured confounder—we control for a variety of other factors that influence justices’ votes. Here, we briefly list these variables and their measurement; the online supplement provides additional detail. First, we control for whether the opposing attorney ever clerked at the Supreme Court. If so, we code Opposing Clerk as 1, and 0 otherwise. We control for Attorney Experience and Opponent Experience by counting the total number of previous cases argued by the attorney and those of his or her opponent (McGuire 1995). We also control for whether the attorney is a Male Attorney (Szmer, Sarver, and Kaheny 2010). We control for whether the attorney was a Petitioner Attorney as well as the attorney’s Net Resource Advantage using the party-specific sliding scale of Collins (2004, 2008). We further account for the attorney’s Ideological Congruence with the Voting Justice (Segal and Spaeth 2002) and whether the Solicitor General Supported the attorney (Black and Owens 2011, 2012). Finally, we control, as best we can, for the underlying quality of the case. We include Lower Dissent, which examines whether a judge in the court below dissented from the ruling against the petitioner. Such a dissent can often indicate that the petitioner’s position is strong. We also include Amicus Support and Amicus Opposition, which account for the number of amicus briefs filed for and against the parties’ positions. Hansford (2004a, 2004b) finds that groups tend to file amicus briefs when they believe their position will win. As such, a number of amici briefs might represent a stronger case for one side or the other.⁵

Preprocessing and Matching

Our interest lies in a “but for” question: but for the experience of being a particular Supreme Court law clerk, would the attorney have won the justice’s vote? To answer this question, we follow a number of studies before ours (e.g., Boyd, Epstein, and Martin 2010) and preprocess our data with matching. The goal of matching is to retain data such that the treatment group is identical to the control group (i.e., the data are balanced), with the only difference being exposure to the treatment—for us, the presence of a former law clerk. The observed differences in outcome identified through subsequent empirical analysis, then, can be attributed to the treatment. In that sense, successful matching is like an experimental design.

We employ coarsened exact matching (CEM; Iacus, King, and Porro 2009, 2011, 2012). CEM addresses the “curse of dimensionality” that plagues one-to-one exact matching (i.e., finding no perfect matches between treatment and control because there are too many variables) by acknowledging that a perfect match between treated and control units is often unnecessary. CEM also enjoys a number of other properties that make it an ideal tool for applied researchers. One risk inherent in analyzing observational data is that the analyst will extrapolate beyond the data actually contained in her sample, something that King and Zeng (2006) label “extreme counterfactuals.” With CEM, “All observations within a coarsened stratum for which we have both a treated and a control unit by definition do not involve extrapolating beyond the data and so these observations will be included; otherwise, they will be removed” (Iacus, King, and Porro 2012, 11). Other work also suggests that CEM is simply better able to “achieve lower levels of imbalance, model dependence, and bias” than other matching methods (King and Nielsen 2019, 8). Another helpful attribute of CEM is its ability to handle multiple treatment effects seamlessly, like those we have in our data.

Preprocessing our data with CEM results in a significant reduction in the levels of imbalance that otherwise would have been present. Prior to matching, the overall level of imbalance is .51, as measured through the approach advanced by Iacus, King, and Porro (2011). (The possible values range between 0, which is no imbalance, and 1, which is complete imbalance.) After matching, it plummets to just .10. This represents, in relative terms, an 81 percent reduction of imbalance within our data. Preprocessing the data through matching thus allows us to remain significantly more confident about the effects of former clerk status than if we had not matched.

Methods and Results

With the matched data in hand, we proceed to estimate a logistic regression model.⁶ We follow the advice of Ho et al. (2007) and include our matching variables as controls. Recall that our question is whether former clerks wield unique influence at the Court, and if so, why. So as not to bury the lede, the data show that former clerks enjoy highly targeted influence as a function of their personal knowledge of the justices for whom they previously worked. In all other situations, however, former clerks are no more successful than those who never clerked at all.

Figure 1 presents results from the underlying logistic regression model. The y-axis shows the likelihood an attorney wins a justice’s vote. The x-axis portrays the four variables we used to measure the attorney’s experience as a former Supreme Court clerk. Consider the leftmost value, which represents an attorney who never clerked on the Court (i.e., Never Clerked). We estimate a .57 [.55, .58] probability that a justice votes for such an attorney.

Figure 1.

Effect of attorney clerk background on justice voting.

Moving one step to the right, consider an otherwise identical attorney who formerly clerked on the Court and who attempts to win the vote of a justice who was not on the Court during that clerkship (i.e., Clerked, Non-sitting Justice). This would be, for example, Mr. Saunders, seeking to win the votes of Justices Sotomayor and Kagan. Attorneys in the same situation as Saunders have the same likelihood of winning their votes—.57 [.55, .59]—as a non-clerk.

We next evaluate a former clerk who attempts to win the vote of a justice who was on the Court during the clerk’s year—but who was not the justice for whom the clerk worked (Clerked, Sitting Justice). For Mr. Saunders in 2015, this would include Chief Justice Roberts and Justices Scalia, Kennedy, Thomas, Breyer, and Alito. Under this scenario, we estimate a former clerk in Mr. Saunders’s situation has a .59 [.58, .61] probability of winning the vote of that justice. Importantly, however, this apparent (minor) advantage is not significantly different from either an attorney who never clerked at all (p = .22) or an attorney who clerked for a justice no longer on the Court (p = .51). In other words, attorneys who clerked while a sitting justice still presides are no more likely to win that justice’s vote than otherwise identical attorneys without clerkship experience or those who clerked for a since-retired justice.

Finally, consider the former clerk who returns to the Court and attempts to win the vote of his or her own justice. For Mr. Saunders, this would be his likelihood of persuading Justice Ginsburg to vote for his client. As the figure makes clear, we find very strong substantive evidence of influence in this context. We estimate a .73 [.71, .74] probability that attorneys like Saunders will be successful—as he ultimately was—at capturing their former justice’s vote. This sixteen-point probability increase represents a 28 percent relative increase compared with someone who never clerked at all and, of course, represents a significant advantage over all three of the other categories of attorneys. To put this value into some perspective, consider the Johnson et al. oral argument study we described earlier. That study finds that quality arguments correlate with a justice’s vote such that a two-standard deviation increase above the average performance gives the attorney a twelve-point probability increase. Our results show that the advantage of being a justice’s former law clerk is fully 33 percent larger than that.

Robustness Checks

In a moment, we will unpack what these findings mean in terms of which theory the data support. Before we do that, however, we pause to consider a number of empirical and modeling concerns that, if left unaddressed, might challenge our results. First, although we use matching to try and best approximate an experimental design with observational data, we must acknowledge our inability to assign attorneys randomly to cases. As a result, we might be concerned about attorney case selection effects. Is there a relationship between whether a former law clerk argues a case and the underlying legal quality of that side’s argument? Perhaps former clerks are better at spotting winnable cases and agree to argue only them. As a preliminary manner, it is worth emphasizing that we already controlled for case quality via three measures: Lower Dissent and Amicus Support/Opposition. Similarly, our empirical results challenge this alternative account. If former clerks systematically cherry-pick cases, then we should find a broad advantage for all former clerks. As Figure 1 shows, however, former clerks are only influential insofar as they can capture their former justices’ votes.

Nevertheless, to explore this concern further, we conducted two auxiliary analyses that leverage novel measures of case strength (only available for a subset of our observations). These include Lindquist and Klein’s (2006) data on legal conflict (the number of circuits lined up on each side of a circuit split) as well as Corley, Steigerwalt, and Ward’s (2013) measure of legal certainty (when the Court produces lopsided or unanimous outcomes). If selection effects explained success, then we should find that former clerks are more likely than non-clerks to sign on to strong cases. For both of these measures, we fail to uncover evidence of a systematic relationship (either positive or negative) between former clerk participation and case strength (see the online supplement for details).⁷

We next consider the role of attorney quality. It is no secret that Supreme Court law clerks are drawn predominantly from the top of the class of top law schools. Clearly, they are a talented group of individuals. It seems reasonable to believe that the same talents that help a person secure a Supreme Court clerkship also help them win cases before the Court. And we do not include specific controls for these factors.⁸ Do our results, then, merely demonstrate that being talented helps you win? We think not. First, our results do not suggest a universal advantage for all former law clerks. Rather, whether a former clerk succeeds is predicated upon having expertise about the justice whose vote she seeks to capture. If quality was a confounding variable, then we would expect to find that former law clerks, by virtue of being the best and the brightest, always outperform non-clerks. They do not.

To probe the strength of this alternative account further, we combined our data with those of Katz and Stafford (2010), who identified nearly every single individual who served as a law clerk to any federal district or circuit court judge from 1995 to 2004. These data allow us to compare the win rate of attorneys who “only” clerked for a circuit court judge as compared with those who clerked for both a circuit court judge and a Supreme Court justice. A circuit court clerkship is coveted, competitive, and, in recent decades, essentially a prerequisite to becoming a Supreme Court law clerk (Baum and Ditslear 2010). If quality is the main driver of Supreme Court clerkship access—and success—before the Court, then we would see differences in the win rates of these two groups of attorneys. We do not.

Katz and Stafford’s data allow us to identify 132 appearances by attorneys with only a circuit court clerkship and 186 appearances by attorneys with both a circuit court and Supreme Court clerkship. The case-level win rates are essentially identical: 60.6 and 60.2 percent, respectively. Differences at the justice vote level are also substantively indistinctive—57.2 percent (circuit only) and 58.9 percent (circuit and Supreme Court)—and statistically insignificant (p = .36, two-tailed test). That these two high-quality groups of attorneys win both cases and votes at virtually the same rate suggests that our results—which we believe derive from personalized information—cannot be explained simply by attorney quality.

Implications for Theories of Revolving Door Influence

We now return to the various theories of former clerk influence. Our results show that former law clerks wield substantial but targeted influence. What leads to this influence? Recall the various theories. The first was that former clerks win because they have substantive expertise in areas of the law that others lack. Our results cast considerable doubt on this explanation. If substantive expertise drove success, we would expect to see broader advantages enjoyed by all former clerks at winning the votes of all justices, as compared to non-clerks. But we do not. Some former clerks are more likely to win than non-clerks. Other former clerks are not.

We are also, with some supplemental analyses, able to reject a more extreme version of the substantive expertise theory. During the terms of our analysis, we know that a number of justices designated one or two of their clerks as the “lead clerks” in specific cases. Justice Blackmun, for instance, required a clerk to prepare a pre-argument bench memo in each case and, if applicable, draft Blackmun’s opinion (Ward and Weiden 2006). Perhaps such clerks become experts in these, more particular, domains. If former clerks systematically gained narrow substantive expertise while clerking and then attempted to leverage that advantage when they returned to argue cases, they would be more likely to win votes in cases where they had such experience.

But they do not. In analyses we report in the online supplement, we looked at whether exposure to a specific issue during one’s time spent clerking translated into an increased likelihood of winning a justice’s vote. We find no relationship between winning and the total number of cases where a clerk’s justice wrote either a majority or separate opinion (p = .98). We do find a systematic relationship between winning and the number of cases the Court heard on an issue during the clerk’s term, but it suggests the opposite is true: greater exposure to an issue is correlated with a decrease in one’s success. Simply put, there is no empirical support for the substantive expertise theory.

We move, next, to the wealth and resources theory. If resources alone led to a greater probability of wining justices’ votes, then former clerks would, across the board, be no more successful than non-clerks. The theory argues that former clerks tend to get hired at more prestigious firms. Indeed, in the unmatched data, there is a pronounced and statistically significant relationship between resource differential and whether an attorney is a former law clerk. Former clerks enjoy between a 1.9- to 2.8-point status advantage compared with attorneys who never clerked. What this means is that if we examined imbalanced data, the resources component could explain former clerk success. Our matching approach, however, deals with that imbalance. In the matched data, the maximum difference is just –.35, which is not statistically significant. In other words, matching removes much of the resource advantage that might otherwise explain the results. Once examining matched data, if party resources alone explained former clerk influence, former clerks would not be more successful than non-clerks. But as Figure 1 shows, some of them are. Resources might go far, and they might explain some former clerk success, but there is much more to the story.

The results also require that we reject the process expertise theory. If process expertise explained former clerk success, all former clerks would perform better than non-clerks. After all, former clerks would know how the system operates. Non-clerks would not. As we discussed already, however, former clerks for non-sitting justices and former clerks for sitting justices are no more likely to win their votes than non-clerks. When we compare these attorneys with otherwise identical advocates who never clerked, we fail to find any significant difference between the two. Simply having clerked and learned how the Court operates internally does not make a former clerk more successful.

We turn, finally, to the personalized information theory. According to this theory, a former clerk knows what his former boss wants to see, knows what makes that justice tick, and can use that personal information to his advantage. There are strong reasons to believe this theory. Clerks are close with their justices. They “have a private dining room in the Court’s cafeteria and attend lunches, receptions, and parties with” their justices (Ward and Weiden 2006, 2). They work closely with their justices in a sort of master–apprentice relationship. Not only do they work with their justices to determine whether they should vote to hear cases (Benesh, Armstrong, and Wallander 2020; Black and Boyd 2012), in many chambers they put together bench memos that their justices will use to ask questions during oral argument (Johnson, Stras, and Black 2014). The clerks are privy to the candid reactions of a justice to all sorts of legal arguments. Chief Justice Rehnquist (2001, 240) put it well: “I tell the clerk some of my reactions to the arguments of the parties, and am interested in getting the clerk’s reactions to these same arguments.” Think back to Mr. Saunders. He spent a year learning to understand law like Justice Ginsburg, to write like Justice Ginsburg, and to think like Justice Ginsburg. All this should come in handy when he later tries to win her vote.

Perhaps not surprisingly, our analysis strongly supports the personalized information theory. Attorneys are most likely to capture their former justices’ votes. Indeed, as we discussed in the context of Figure 1, clerks for employing justices are always more likely to capture their justice’s votes than other former clerks and non-clerks. More specifically, justices are 16 percentage points more likely to vote for their former clerks than for otherwise identical non-clerk. They are 16 percentage points more likely to vote for their former clerks than someone who clerked for a since-retired justice. And, perhaps most telling, they are 14 percentage points more likely to vote for their former clerks than someone who clerked for another one of their colleagues. This seems to us a substantial advantage.

One alternative reading of these empirical results focuses not on what the attorney knows about a justice but rather what a justice knows about the attorney. Instead of Saunders knowing the ins and outs of how Justice Ginsburg thinks, perhaps it is that she knows and trusts Saunders and as a result, is more inclined to support his position in a case. To evaluate this possibility, we conducted an auxiliary analysis examining whether the relationship we document above is conditioned by the number of appearances an attorney made before a voting justice. If trust is the driving component, repeated exposure to an attorney should act to attenuate the disadvantage some attorneys face. It does not, however. Non-clerks exhibit the same likelihood of winning a justice’s vote regardless of whether it is their first or fifth appearance before that justice, as do all variants of former law clerks (see the online supplement for details). Simple familiarity does not explain our results.

Practical and Normative Implications

We show that a former law clerk is significantly advantaged relative to all other attorneys when she appears before her former justice. Although such attorney–justice vote dyads historically represent a small percent of observations in the data, there are a number of reasons to appreciate the importance of the finding. For starters, former law clerks are increasingly active before the Court. Figure 2 illustrates. Each of the four panels corresponds to a type of attorney appearing before the Court. The x-axis shows each of the forty terms included in our analysis. The y-axis portrays the percent of cases per term where an attorney of each type participated in a case. The top left panel shows the marked decline of cases with no former clerks. The remaining panels chart the meteoric rise in the various forms of participation by former clerks.⁹ Of particular interest is the fact that in recent terms just as many cases featured former clerks arguing before their justices as cases without them.

Figure 2.

Relative frequency of former law clerk participation across time (1979–2018).

What is more, given the demographics of the current justices, we have little reason to believe these trends will do anything but continue. The median age of Supreme Court justices is currently sixty-five. Because justices do not retire until an average age of nearly eighty years old, we can expect them to cycle through many more cohorts of clerks who will then have plenty of time to return and argue before their former justices. In other words, there is ample reason to believe our findings could become even more pronounced with time.

Equally noteworthy, the Court decides a number of cases by the thinnest of vote margins. Switching even one justice’s vote would flip the outcome in these 5-4 cases. In our data, we observe a total of 710 cases decided by a single vote. Of these 710 cases, fully 10 percent featured a former law clerk who argued before her justice and ultimately ended up winning the 5-4 case and her justice’s vote. Perhaps more tellingly, fully 52 percent of these 5-4 cases featured no former Supreme Court law clerks arguing for the losing side.¹⁰ Did a justice’s former clerk sway the winning vote in the former set of cases? Would a former clerk have made a difference in the latter group? Making these sorts of claims is beyond what our data allow, but the broader point these numbers illustrate is that our results have sizable potential implications for legal policy in the United States.

We suspect one other point is worth highlighting. On the modern Court, the median justice tends to get what he or she wants. Recently, Court watchers referred to the “O’Connor Court” or to the “Kennedy Court.” This label, of course, reflects the power of the median justice. If the median justice is so powerful, and if clerks have the most influence over their former justices, we suspect that former clerks for the median justice will be in high demand. The stock of former Kennedy clerks, by our accounting, should have dropped when he announced his retirement. On the contrary, former clerks for Chief Justice Roberts are likely in high demand now that he is the median justice. (Clerks for soon-to-be retired justices, like Ginsburg and Breyer, may not be worth the same investment as clerks whose former justices have potential to become the median.)

These results raise concerns about fairness. The fact that an attorney may be more likely to win a justice’s vote because of personal familiarity seems problematic, particularly when one considers that obtaining that familiarity is, itself, a highly selective affair. Our analyses show that former clerks benefit from personalized information about their justices. But what does it take to become a Supreme Court clerk? Obviously, it requires skill and intelligence. At the same time, however, the fact that most justices draw their law clerks from only a handful of law schools (Harvard, Yale, Stanford, Columbia) suggests that an attorney who does not go to one of those schools will be significantly disadvantaged down the road when making arguments to the Court. That is, the disadvantage of not being a former clerk can set in years before the attorney even appears before the Court. Justices might wish, therefore, to consider following the example of Justice Thomas and hire equally talented clerks from law schools outside the traditional elite schools.

Conclusion

We began this study by pointing out the huge sums of money law firms throw at recently minted Supreme Court clerks. One might ask whether it is worth it for them to do so. Our results show that the answer is yes. Consider that not-so-rare circumstance where profound financial consequences are associated with winning or losing the case. As the head of appellate litigation at one D.C. law firm stated, “if you’re working on billion-dollar cases, the client is willing to pay more for truly excellent work” (Lithwick 2007). We suspect many clients would leap at the opportunity to influence a justice by any value, let alone one that moved sixteen full points.

Former Supreme Court law clerks have many advantages: they are highly skilled, intelligent, and well resourced. These factors surely influence their success rates when they appear later before the High Court. But there is something else that sets them apart from other attorneys: the personalized information they hold about the justices for whom they worked. The data show that attorneys who know much about a justice are more likely to capture that justice’s vote.

Looking more broadly at revolving door lobbying, the literature from which we drew our theories, the results suggest that such lobbying success might also be a function, in large part, of personalized information. To be clear, Supreme Court litigation is not the same thing as lobbying. The two activities surely involve different behaviors, different actors, and different expectations. Nevertheless, the findings are suggestive. It could very well be that revolving door lobbyists succeed because of their personal knowledge of relevant decision makers.

In Peugh v. United States (2013), the Court examined whether Marvin Peugh should have been sentenced under the Sentencing Guidelines in effect at the time he was convicted of bank fraud or under the Guidelines as they had been revised by the time he got to sentencing. The Court ruled 5-4 that to use the newer Guidelines violated the Ex Post Facto clause of the constitution. Justice Kennedy voted with the four liberal justices. His vote was somewhat surprising, as it was the only liberal vote Kennedy ever cast in bank robbery cases. Perhaps it ought not to have been surprising though. It was his former clerk, Stephen Kinnaird, who argued on behalf of Peugh.

Supplemental Material

online_appendix – Supplemental material for The Influence of Personalized Knowledge at the Supreme Court: How (Some) Former Law Clerks Have the Inside Track

Supplemental material, online_appendix for The Influence of Personalized Knowledge at the Supreme Court: How (Some) Former Law Clerks Have the Inside Track by Ryan C. Black and Ryan J. Owens in Political Research Quarterly

Footnotes

Acknowledgements

For providing highly helpful assistance with this project, we gratefully thank William Grant, Eli Pales, and Pernod Ricard.

Authors’ Note

An earlier version of this paper was presented at the 2016 Annual Meetings of the Southern Political Science Association (January 7–9 in San Juan, Puerto Rico).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ryan C. Black

Supplemental Material

Files to replicate the empirical results of this paper and the online supplement can be found on Harvard Dataverse: https://doi.org/10.7910/DVN/JOCRPU. The online supplement, mentioned at various points in the article, is available at .

Notes

References

Bailey

M. A.

Kamoie

Maltzman

2005. “Signals from the Tenth Justice: The Political Role of the Solicitor General in Supreme Court Decision Making.” American Journal of Political Science 49:72–85.

Baum

Ditslear

2010. “Supreme Court Clerkships and ‘Feeder Judges.’” Justice System Journal 31:26–48.

Baumgartner

F. R.

Berry

J. M.

Hojnacki

Kimball

D. C.

Leech

B. L.

2009. Lobbying and Policy Change: Who Wins, Who Loses, and Why. Chicago: University of Chicago Press.

Benesh

S. C.

Armstrong

D. A.

Wallander

2020. “Advisors to Elites: Untangling Their Effect.” Journal of Law and Courts 8:51–73.

Black

R. C.

Boyd

C. L.

2012. “The Role of Law Clerks in the U.S. Supreme Court’s Agenda-Setting Process.” American Politics Research 40:147–73.

Black

R. C.

Hall

M. E. K.

Owens

R. J.

Ringsmuth

E. M.

2016. “The Role of Emotional Language in Briefs before the U.S. Supreme Court.” Journal of Law and Courts 4:377–407.

Black

R. C.

Owens

Ryan J.

2009. “Agenda-Setting in the Supreme Court: The Collision of Policy and Jurisprudence.” Journal of Politics 71(3): 1062–1075.

Black

R. C.

Owens

R. J.

2011. “Solicitor General Influence and Agenda Setting on the U.S. Supreme Court.” Political Research Quarterly 64:765–78.

Black

R. C.

Owens

R. J.

2012. The Solicitor General and the United States Supreme Court: Executive Influence and Judicial Decisions. New York: Cambridge University Press.

10.

Boyd

C. L.

Epstein

Martin

A. D.

2010. “Untangling the Causal Effects of Sex on Judging.” American Journal of Political Science 54:389–411.

11.

Cain

B. E.

Drutman

2014. “Congressional Staff and the Revolving Door: The Impact of Regulatory Change.” Election Law Journal: Rules, Politics, and Policy 13:27–44.

12.

Collins

P. M.

Jr.

2004. “Friends of the Court: Examining the Influence of Amicus Curiae Participation in U.S. Supreme Court Litigation.” Law & Society Review 38:807–32.

13.

Collins

P. M.

Jr.

2008. Friends of the Court: Interest Groups and Judicial Decision Making. New York: Oxford University Press.

14.

Corley

P. C.

2008. “The Supreme Court and Opinion Content: The Influence of Parties’ Briefs.” Political Research Quarterly 61:468–78.

15.

Corley

P. C.

Steigerwalt

Ward

2013. The Puzzle of Unanimity: Consensus on the United States Supreme Court. Stanford: Stanford University Press.

16.

Drutman

2014. “Interest Groups.” In Developments in American Politics, edited by Peele

Bailey

C. J.

Cain

B. E.

Peters

B. G.

, 7th ed., 75–92. London: Palgrave Macmillan.

17.

Esterling

K. M.

2004. The Political Economy of Expertise. Ann Arbor: University of Michigan Press.

18.

Hanmer

M. J.

Kalkan

K. O.

2013. “Behind the Curve: Clarifying the Best Approach to Calculating Predicted Probabilities and Marginal Effects from Limited Dependent Variables.” American Journal of Political Science 57:263–77.

19.

Hansford

T. G.

2004a. “Information Provision, Organizational Constraints, and the Decision to Submit an Amicus Curiae Brief in a U.S. Supreme Court Case.” Political Research Quarterly 57:219–30.

20.

Hansford

T. G.

2004b. “Lobbying Strategies, Venue Selection, and Organized Interest Involvement at the U.S. Supreme Court.” American Politics Research 32:170–97.

21.

D. E.

Imai

King

Stuart

E. A.

2007. “Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference.” Political Analysis 15:199–236.

22.

Iacus

S. M.

King

Porro

2009. “CEM: Software for Coarsened Exact Matching.” Journal of Statistical Software 30:1–27.

23.

Iacus

S. M.

King

Porro

2011. “Multivariate Matching Methods that Are Monotonic Imbalance Bounding.” Journal of the American Statistical Association 106:345–61.

24.

Iacus

S. M.

King

Porro

2012. “Causal Inferences without Balance Checking: Coarsened Exact Matching.” Political Analysis 20:1–24.

25.

Johnson

T. R.

Stras

D. R.

Black

R. C.

2014. “Advice from the Bench (Memo): Clerk Influence on Supreme Court Oral Arguments.” Marquette University Law Review 98:21–42.

26.

Johnson

T. R.

Wahlbeck

P. J.

Spriggs

J. F.

II . 2006. “The Influence of Oral Argumentation before the U.S. Supreme Court.” American Political Science Review 100:99–113.

27.

Katz

D. M.

Stafford

2010. “Hustle and Flow: A Social Network Analysis of the American Federal Judiciary.” Ohio State Law Journal 71:467–510.

28.

Kendall

2012. “High-Court Clerks Attract a Frenzy.” The Wall Street Journal, September 18. https://www.wsj.com/articles/SB10000872396390443720204578000202789898248.

29.

King

Nielsen

2019. “Why Propensity Scores Should Not Be Used for Matching.” Political Analysis 27:435–54.

30.

King

Zeng

2006. “The Dangers of Extreme Counterfactuals.” Political Analysis 14:131–59.

31.

LaPira

T. M.

Thomas

H. F.

2017. Revolving Door Lobbying: Public Service, Private Influence, and the Unequal Representation of Interests. Lawrence: University Press of Kansas.

32.

Lat

2015. “Which Law Firm Won the SCOTUS Clerk Sweepstakes?” Above the Law, November 2. https://abovethelaw.com/2015/11/which-law-firm-won-the-scotus-clerk-sweepstakes-you-should-know-by-now/.

33.

Lazarus

R. J.

2008. “Advocacy Matters Before and Within the Supreme Court: Transforming the Court by Transforming the Bar.” The Georgetown Law Journal 96:1487–1564.

34.

Lindquist

S. A.

Klein

D. E.

2006. “The Influence of Jurisprudential Considerations on Supreme Court Decisionmaking: A Study of Conflict Cases.” Law & Society Review 40:135–62.

35.

Lithwick

2007. “What to Make of Those Astronomical Supreme Court Signing Bonuses?” Slate, March 10. https://slate.com/news-and-politics/2007/03/what-to-make-of-those-astronomical-supreme-court-signing-bonuses.html.

36.

Long

L. N.

Christensen

W. F.

2008. “Clearly, Using Intensifiers Is Very Bad, or Is It?” Idaho Law Review 45:171–89.

37.

Mauro

2018. “Jones Day Lands a Record 11 Supreme Court Law Clerks as Associates.” The National Law Journal, November 13. https://www.law.com/nationallawjournal/2018/11/13/jones-day-lands-a-record-11-supreme-court-law-clerks-as-associates/?slreturn=20200622055413.

38.

McGuire

K. T.

1993a. “Lawyers and the U.S. Supreme Court: The Washington Community and Legal Elites.” American Journal of Political Science 37:365–90.

39.

McGuire

K. T.

1993b. The Supreme Court Bar: Legal Elites in the Washington Community. Charlottesville: University of Virginia Press.

40.

McGuire

K. T.

1995. “Repeat Players in the Supreme Court: The Role of Experienced Lawyers in Litigation Success.” Journal of Politics 57:187–96.

41.

McGuire

K. T.

2000. “Lobbyists, Revolving Doors and the U.S. Supreme Court.” Journal of Law and Politics 16:113–37.

42.

Miller

Keith

L. C.

Holmes

J. S.

2015. “Leveling the Odds: The Effect of Quality Legal Representation in Cases of Asymmetrical Capability.” Law and Society Review 49:209–39.

43.

O’Connor

Hermann

J. R.

1995. “The Clerk Connection: Appearances before the Supreme Court by Former Law Clerks.” Judicature 78:247–49.

44.

Peppers

T. C.

2006. Courtiers of the Marble Palace: The Rise and Influence of the Supreme Court Law Clerk. Stanford: Stanford University Press.

45.

Rehnquist

W. H.

2001. The Supreme Court. Revised and Updated. New York: Vintage Books.

46.

Schoenherr

J. A.

Black

R. C.

2019. “The Use of Precedent in U.S. Supreme Court Litigant Briefs.” In Research Handbook on Law and Courts, edited by Sterett

Walker

L. D.

, 114–29. Cheltenham: Edward Elgar.

47.

Segal

J. A.

Spaeth

H. J.

2002. The Supreme Court and the Attitudinal Model Revisited. New York: Cambridge University Press.

48.

Szmer

Ginn

M. H.

2014. “Examining the Effects of Information, Attorney Capability, and Amicus Participation on U.S. Supreme Court Decision Making.” American Politics Research 42:441–71.

49.

Szmer

Johnson

S. W.

Sarver

T. A.

2007. “Does the Lawyer Matter? Influencing Outcomes on the Supreme Court of Canada.” Law and Society Review 41:279–304.

50.

Szmer

Sarver

T. A.

Kaheny

E. B.

2010. “Have We Come a Long Way, Baby? The Influence of Attorney Gender on Supreme Court Decision Making.” Politics and Gender 6:1–36.

51.

Vidal

J. B.

Draca

Fons-Rosen

2012. “Revolving Door Lobbyists.” American Economic Review 102:3731–48.

52.

Ward

Dwyer

Gill

2014. “Bonus Babies Escape Golden Handcuffs: How Money and Politics Has Transformed the Career Paths of Supreme Court Law Clerks.” Marquette University Law Review 98:227–59.

53.

Ward

Weiden

D. L.

2006. Sorcerers’ Apprentices: 100 Years of Law Clerks at the United States Supreme Court. New York: New York University Press.

54.

Wedeking

J. P.

2010. “Supreme Court Litigants and Strategic Framing.” American Journal of Political Science 54:617–31.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.16 MB