Vol. 28 No. 4 (July 2018) pp. 45-51ETHICS AND ACCOUNTABILITY ON THE U.S. SUPREME COURT: AN ANALYSIS OF RECUSAL PRACTICES, by Robert J. Hume. Albany, NY: State University of New York Press, 2017. 204pp. Paper $20.95. ISBN: 9781438466965.
Reviewed by Ryan C. Black, Department of Political Science, Michigan State University. Email: rcblack@msu.edu.
Scholars of the U.S. Supreme Court commonly observe that one of the most important powers of the Court is its near-universal control over its docket. As the title to H.W. Perry’s seminal volume on agenda-setting reminds us (1991), the Court has the power to decide what it will decide. In ETHICS AND ACCOUNTABILITY ON THE U.S. SUPREME COURT, Robert Hume contends that the decision to recuse (or not) from a case provides Supreme Court justices with even more discretion than is commonly acknowledged. This book is ambitious in that it seeks “to penetrate the myths surrounding recusals by studying their causes and consequences systematically” (p. 4). Hume’s efforts are overwhelmingly successful and the result is a compact and accessible text that significantly enhances our understanding of an aspect of decision making that lands in the scholarly sweet spot of being really important but woefully understudied.
As a preliminary matter, Hume makes a persuasive case for why recusal is worthy of scholarly attention. First, the identity of who votes in cases has profound implications for both the disposition of cases and the resulting content of law. When a justice sits out from a case, what was once a comfortable majority might become more tenuous. Secondly, this dynamic is especially important at the agenda-setting stage, where the Court’s “Rule of Four” requires at least four votes – out of nine – to grant review to a petition. This bar does not lower in the absence of one (or more) justice. Third, beyond individual cases or petitions, discretionary recusals provide a uniquely clear window into how justices and the Court seek to balance a number of competing interests such as institutional legitimacy and fairness versus the desire to set legal policy.
Chapter 1 combines a pithy discussion of a handful of salient recusal decisions (or non-decisions) with novel empirical data on media coverage to bring recusal novices up to speed. Prior to the 1970s, the limited amount of editorial commentary on recusals opposed their usage for concern that their usage would create administrative problems for the Court (e.g., evenly-divided courts). Discussion of Judge Clement Haynsworth’s failure to recuse himself as a circuit court judge provided some of the first fodder for calling for more usage of recusals. Around the same time, then Associate Justice William Rehnquist found himself in the editorial crosshairs over his non-recusal in a trio of cases involving the Nixon administration while he was Assistant Attorney General. This culminated in what was the first public statement issued by a justice defending their participation in case.
Since the 1970s there have been two salient flare-ups concerning recusal decisions. The first, in 2004, involved Justice Antonin Scalia issuing a statement justifying his participation in a case involving Vice President Dick Cheney after it was revealed that Scalia had spent a weekend duck hunting with a group of individuals that included Cheney. Quick (and interesting) tangent: Hume reports on the existence of earlier, arguably more egregious, instances of cross-branch fraternization during the 20th century that were previously unknown to this reviewer. Justices Robert Jackson and Byron White vacationed with President Franklin Roosevelt and Attorney General Bobby Kennedy, respectively, while cases involving their administrations were pending before the Court (p. 5-6).
The second recusal flare-up took place when the Court was poised to review the constitutionality [*46] of the Affordable Care Act. In this case, Justice Elena Kagan’s participation was criticized by some, given her role as the Solicitor General under President Obama (though Kagan maintains that she was uninvolved in related discussions). Similarly, Justice Clarence Thomas’ participation was scrutinized not for his behavior, but rather that of his wife, Ginny Thomas, who was a vocal Obamacare opponent.
Chapter 2 provides setup for an empirical analysis of recusal behavior. Here, Hume identifies three motivations that guide recusals. First, there is the concern of compliance with the statutory language that identifies the conditions under which justices should recuse. Having a financial stake in the case is thought to be the most inviolable of these, and this most commonly manifests itself when a justice’s investment portfolio includes stock of a company party to the case. Other disqualifying conditions include situations where a judge was previously involved in the proceeding or has expressed an opinion concerning the merits of the specific case, or a family member or close personal relationship is involved in a case. The role of these four factors is to identify instances where the “right” thing to do is not participate in a case.
What makes recusal institutionally unique on the Supreme Court is the inability, at present, for a recused justice to be replaced by another jurist – such a consideration does not exist on the lower courts, where other judges can (and do) fill in for those who are recused. The Court’s role as the final arbiter on legal matters in the U.S. further heightens the stakes when it comes to risking an evenly-divided court. The justices subsequently tend to maintain the view that they have a “duty to sit” whenever possible.
Lastly, it should come as no surprise that, as a political scientist, Hume also identifies a policy-based component to the recusal decision. However, his approach is more nuanced and subtle than the typical “single-minded seeker of policy” shtick that we all too often see. Rather, he quite cleverly argues that a justice’s desire to implement policy promotes a confirmation bias whereby they “devalue evidence that would seem to favor recusal when they have strong policy reasons for participating” (p. 34). Justices, then, can still believe they are behaving ethically even if they might make choices that are on the edge of appropriateness. I found this argument refreshing, especially as my reading of the book so far had led me to think (perhaps wrongly) that Hume was more cynical regarding justices’ motivations.
Chapter 3, which examines the conditions under which justices recuse themselves, is the first of several empirical chapters in the volume. From his statutory, institutional, and policy motivations, Hume identifies a total of thirteen hypotheses. Data for his analyses come from the venerable Supreme Court Database, which he supplements with original data. The main analysis in this chapter examines over 73,000 justice recusal decisions spanning from 1946-2010. Hume finds support for elements of all three of the aforementioned motivations. In terms of compliance with recusal statutes, justices are more likely to recuse when a business or similar entity is involved in the case (i.e., a justice might have a possible financial stake in the case). Similarly, a justice who formerly served as the Solicitor General (SG) and was likely involved in a good number of the cases before the Court is also more likely to recuse themselves than a justice who was not the SG. This effect is, predictably, attenuated by the passage of time and a new batch of cases working its way through the system. One unexpected result on the statutory side of things is Hume’s finding that justices who came to the Court by way of serving on a federal circuit court are actually less likely to recuse than those who did not.
As for institutional considerations, Hume’s analysis reveals the importance of two hypothesized factors. First, cases decided on statutory interpretation grounds are more likely to see recusal than those decided on constitutional grounds. Secondly, cases granted review because of legal conflict are less likely to see a justice recuse than those that are not. The substantive magnitude of these effects is not overwhelming, but this is to be expected given [*47] the low probability of observing recusals at all (overall about 2% in Hume’s data).
Last, and the motivation likely to raise the most eyebrows among readers, is policy motivations. Hume hypothesizes that a justice’s ideological distance from the median justice matters, but in a rather sophisticated way. He argues that justices who are ideologically proximate to the median should be less likely to recuse as so doing would cause them to give up their ability to exert meaningful influence on the case outcome and resulting opinion. As distance from the median increases, recusal should be more likely as the policy incentive decreases (i.e., the justice becomes less pivotal to the case outcome), but then, for the highest values of ideological distance, policy incentives return and recusal likelihood reduces again. This happens, Hume persuasively argues, because the most distant justices are motivated to stay involved in the case so as to provide the countervailing perspective (likely in the dissent) or otherwise work to keep the majority opinion honest.
And, drumroll please, it works! Mostly. The justices who are least likely to recuse themselves are those who are in roughly the 70th percentile of distance from the median. As one gets farther from that level, the recusal probability increases modestly until the most extreme values in the data (i.e., 95th+ percentile), at which point the probability decreases again. My “mostly” modifier refers to what happens with recusal probability as you work from very low values of distance up to the 70th percentile value. In particular, the recusal probability is among the highest for when a justice is the median (i.e., when distance is equal to zero). This strikes me as strange since the median arguably has the most policy influence and should be, I think, the least likely to want to give that up. Similarly, as you start to walk slowly away from the median to being more ideologically distant, the probability of recusal actually decreases (as opposed to increases). These justices, it seems to me, should have less to gain ideologically from their participation, so it is unclear to me why they’re still participating if Hume’s hypothesis is true.
The chapter concludes with a deeper dive into the financial conflict of the interest aspect of recusal decision making. Using the justice’s financial disclosure documents (only available from 1980-2010), Hume re-estimates his earlier model and reports a null relationship between a justice’s net worth and whether he or she recuses in a case. However, with a t-statistic of -1.57, the results actually approach some level of statistical significance in the opposite direction, with justices who have higher net worth in a reported year being less likely to recuse. It is unclear why this reverse finding might exist. Similarly, I also wonder if a more appropriate test of this hypothesis would be to interact a justice’s net worth with whether the parties in a case are business or similar entities. As Hume notes, justices with higher net worth and holdings should have more conflicts (p. 61), but a recusal-worthy financial conflict seems likely to exist only if some sort of business entity is party to the case.
Chapter 4 takes up the topic of the policy consequences of Supreme Court recusals with three related dependent variables. First, Hume examines whether a recusal is correlated with a shift in the ideological direction of the Court’s decisions. Additionally, he examines if a recusal affects the ideological location of a case’s majority opinion author and, third, he looks at the ideological position of the median of the Court’s majority coalition. To assess impact, he takes a case-centered approach and operationalizes recusal via two measures: whether a recused justice is liberal and whether a recused justice is conservative. Each of these is a dummy variable coded as 1 if a recused justice is to the left or right of the Court’s median, respectively.
Hume finds that a conservative recusal significantly reduces the probability that the Court reaches a conservative case outcome. As Hume is quick to note, however, this effect appears only in two types of cases: those that are not in the civil liberties issue area and those that were subsequently not salient after being decided (via Epstein and Segal’s [2000] New York Times measure). Stated a bit differently, in terms of case outcomes, the absence of a [*48] conservative justice has no significant impact in civil liberties or salient cases. The recusal of a liberal justice is not a significant influence on case outcomes for any type of case.
Unlike case outcomes, Hume reports no significant impact in terms of the ideology of the majority opinion writer, writing that, “One can therefore safely conclude that recusals do not have any direct impact on opinion assignments” (p. 82). This statement is, given the results, a bit too strong for my tastes. As before, a number of the results are in the neighborhood of being statistically significant – enough so that I’d be worried about committing a Type II error in accepting the null hypothesis. To wit, there is modest evidence that a liberal recusal actually shifts the majority opinion author ideology to the left and that a conservative recusal similarly pushes the majority opinion author ideology to the right in media-salient cases. This would suggest that the majority opinion assigner perhaps reacts to the loss of an ideological ally by trying to overcompensate in choosing an assignee to write in a case.
Finally, Hume reports what he deems to be the most robust evidence of recusal policy impact on of the ideology of the median member of the majority coalition (MOMC). In particular, a conservative justice’s recusal – across all cases and sub-cases – results in a leftward shift in MOMC ideology. A liberal recusal similarly translates into a conservative shift of approximately the same substantive magnitude. For Hume, this is a noteworthy finding given that the Court’s opinion “describes the legal rule and establishes the precedent for lower courts to apply. The way an opinion is written therefore has strong potential to influence the future development of the law” (p. 82, 87).
I am, for several reasons, less convinced. Although I have no doubt that the MOMC matters, they aren’t going to be the justice who determines or constrains how far the majority is able to go. That power lies with the justice who is ideologically the closest to the cutpoint between the majority and dissent coalitions. If the majority goes too far, this justice at the edge is the one who could weaken (or break) the coalition by defecting to the other side.
Secondly, I see a shift in MOMC ideology as being a necessary but not sufficient condition to say that legal content has changed. Imagine the most common Supreme Court case, which in the terms analyzed by Hume will yield a 9-0 outcome. Now randomly recuse one justice. The ideological location of the MOMC will, necessarily, move in the way found by Hume’s analysis. But the decision would still be unanimous and my hunch is that the opinion is still going to establish the same legal rule it would but for the absence of the recused justice. Without a finer-grain measure of opinion content, it is difficult to say that recusal really results in an appreciable change in the Court’s policy output. However, I think these concerns actually support the general argument Hume wants to make, which is that the so-called policy consequences of justices recusing themselves are not especially noteworthy.
Chapter 5, the final empirical chapter of the book, probes the other consequences of recusals. Hume examines the impact of recusal on evenly-divided courts, public attitudes towards the Court, bargaining that takes place on the Court, and how the Court sets its discretionary agenda. This is an impressive chapter in terms of the scope of data Hume brings to bear on the question. If anything, I was a bit disappointed that he didn’t split them out into meatier individual chapters. But, that’s small potatoes. Evenly-divided courts pose arguably the biggest legal risk to recusal, since their occurrence can preserve ambiguity or conflict in the law. Hume finds that although tie votes are, by definition, more likely to occur when the Court sits with an even number of justices, the substantive magnitude of their risk is generally limited to one or two percentage points.
Hume next looks at what impact recusal might have on public attitudes towards the Supreme Court. This is, for my money, the most important aspect of the recusal debate. Justices and the Court more generally have been remarkably adept at avoiding the widespread belief that they are just politicians in robes. If, [*49] however, it was perceived that they are playing fast and loose with their ethical obligations, then this belief could be challenged. Hume marshals existing as well as novel survey data to assess this concern. A comparison of before and after the Scalia-Cheney duck hunting controversy shows no significant decrease in the percentage of people having “a great deal” or “quite a lot of” of confidence in the Supreme Court. Interestingly, the percentage who has “very little” or “none” does nudge up between the two survey dates (June 2003 and May 2004) from 13 percent to 16 percent (a 23 percent relative increase), but given the myriad things taking place during that intervening year (such as an entire term’s worth of decisions!), Hume is wisely unwilling to assign blame to the recusal controversy.
In addition to the existing data, Hume also reports the results of a novel online survey he fielded in spring 2016 after the unexpected death of Justice Scalia. When asked about attitudes toward the Court dividing equally versus justices compromising their preferences, a majority (about 54 percent) approved of equal divisions compared to just around 29 percent who disapproved. Hume analyzes how one’s response to this question is correlated with an individual’s approval of the Court and legitimacy attitudes as well, finding positive and null effects, respectively.
I must confess to being somewhat disappointed by this aspect of the chapter. The key limitation of the public confidence survey data presented earlier in the chapter was their inability to attribute causation to the Scalia-Cheney recusal controversy. The virtue of “rolling your own” survey is that the researcher can randomly manipulate the questions or text a respondent sees and then measure how that manipulation causes subsequent attitudes. This is precisely the approach taken by Gibson and Caldeira’s (2012) important analysis of recusals’ ability to restore public faith in courts where the judges campaign for office and receive contributions or support (short answer: they can, but not entirely). In contrast, it is impossible to determine whether support for even divisions causes approval of the Court or if the causal arrow is reversed. Hopefully Hume has a sequel in the works where he’ll pick up the money that he’s left on the table by examining the recusal of Supreme Court justices with a more rigorous experimental approach.
Turning to the remainder of the empirical evidence examined, Hume cleverly taps into the Burger Court Opinion Writing Database to examine how recusal influences bargaining and accommodating on the Court. Here, he finds that when the majority is a minimum winning coalition (i.e., just five votes), the recusal of a conservative justice leads the majority opinion writer to circulate significantly more drafts than when the Court is operating with a full contingent of justices. This finding is interesting – and important – as it provides what I believe to be the first systematic evidence that the Court consciously works to avoid even divisions when deciding a case with only eight justices.
Last, Hume applies similar scrutiny to the agenda-setting stage and examines the docket sheets of Justice Harry Blackmun to determine what, if any, impact recusal had on the cases granted review by the Court from its 1986-1993 terms. He concludes that in only two instances did recusal have a “but for” impact on the Court’s decision to hear a case. As he wryly observes, the Court disposed of nearly 43,000 cases during that period of time, making the substantive impact of recusal so trivially small as to be essentially zero.
Chapter 6 concludes with Hume’s assessment of proposals for recusal policy reform. Here, Hume is generally skeptical of the need for these at all. As he notes, “on balance, the arguments in favor of reform are not compelling. After reviewing the data, I am led to conclude that the potential benefits of reform are too modest and come at the risk of needlessly burdening the Court” (p. 117). Ever the good sport, Hume proceeds onward to describe three general types of possible reform. First, Congress could seek to impose greater oversight on the process. Presently, authority over the participation decision rests solely with an individual justice. Instead, Congress could attempt to vest either some external body or the Supreme Court itself [*50] with the power to review these decisions. Hume is appropriately skeptical of the prospects for either of these possibilities. It is unclear, for instance, whether the former would survive constitutional review (to be performed, of course, by the Supreme Court). As for the latter, the importance of the norm of collegiality within the Court makes it likely that any review would be perfunctory at best and unlikely to generate substantively different outcomes.
A second type of proposal seeks to address the main downstream consequence of recusal: the possibility of an evenly-divided court. The duty-to-sit doctrine is primarily motivated by the fact that no one can replace a missing justice. Reforms in this vein would seek to correct for this by allowing, for instance, retired justices to fill a temporary vacancy created by recusal. Such a proposal was offered by Senator Patrick Leahy in 2010 and came, reportedly, after a meeting between the senator and Justice John Paul Stevens. Hume expresses concern that such a policy “might discourage recusals because the justices would have greater uncertainty about how substitute judges would resolve disputes in their absence” (p. 129). This concern is actually one that I know a bit about and can happily assuage Hume’s concerns. Amanda Bryan and I conducted a comprehensive empirical assessment of the use of judicial replacements on the Supreme Court (Black and Bryan 2014). We find, among other things, that even under the assumption that justices care about policy and policy alone, the remaining members of the Court would be able to arrive at a consensus replacement pick in over 60 percent of the instances where a vacancy exists.
The final type of reform is one where the justices enhance transparency by adhering to stricter reporting requirements, thus adding clarity to the otherwise murky process by which recusal decisions are made. Such a reform would ultimately build a body of precedents that justices could refer to when trying to decide whether or not to participate in a case. Proponents of this reform argue that transparency “could help justices to maintain public confidence in the institution” (p. 130), but I tend to side with the critics, who believe legitimacy could be damaged by “putting judicial ethics constantly on the public agenda” (p. 130). My view is due in no small part to Gibson and Caldeira’s (2012) finding from the aforementioned recusal analysis that public attitudes toward state judicial candidates are harmed not only if they accept campaign contributions but even when they are offered and reject such contributions. In short, the public seems likely to assess the Court as guilty by the mere implication that their decision making might deviate from being impartial.
One interesting possibility for future work would be to empirically assess if these reforms do anything to alleviate public concern over Supreme Court justices’ conflicts of interest. That is, experimentally manipulate the reform that individuals learn about to see whether confidence, support, or legitimacy attitudes are enhanced as compared to the status quo. The debate over recusal policy seems unlikely to go away anytime soon, so scholars should seek to provide robust empirical contributions to help guide this important discussion.
All told, Hume has provided the scholarly community with a fascinating initial assessment of an under-appreciated yet undeniably important aspect of Supreme Court decision making. As the depth of my above comments and (what I hope will be seen as) friendly critiques reveal, it was a text I greatly enjoyed reading and thinking deeply about. Unlike a lot of scholarly volumes, I do think there is a possible place for this in the classroom as well. It is accessible enough for advanced undergraduates (perhaps with a bit of statistical handholding) and could be used to stimulate discussion and debate in a course on the U.S. Supreme Court, for instance. At the very least, it’s a book that anyone who regularly teaches or researches on the Supreme Court needs to add to his or her bookshelf.
REFERENCES:
Black, Ryan C. and Amanda C. Bryan. 2014. “Calling in the Reserves: Judicial Replacements on the U.S. Supreme Court.” JUSTICE SYSTEM JOURNAL 35(1): 4-26.
[*51] Gibson, James L. and Gregory A. Caldeira. 2012. “Campaign Support, Conflicts of Interest, and Judicial Impartiality: Can Recusals Rescue the Legitimacy of Courts?” JOURNAL OF POLITICS 74(1): 18-34.
Perry, H.W. 1991. DECIDING TO DECIDE: AGENDA SETTING IN THE UNITED STATES SUPREME COURT. Cambridge, MA: Harvard University Press.
© Copyright 2018 by author, Ryan C. Black.
