A. Conceptualizing International Courts

Recent years have witnessed a dramatic judicialization of international relations. In 2016 alone, international courts and arbitral panels issued judgments and awards convicting Radovan Karadžić, wartime leader of the Bosnian Serbs, for genocide;Footnote ⁶ rejecting Philip Morris's claim that a Uruguayan law governing the packaging of cigarettes, designed to promote public health, violated a bilateral investment treaty;Footnote ⁷ dismissing the Marshall Islands' claim that India, Pakistan, and the United Kingdom violated their obligations concerning negotiations over nuclear disarmament;Footnote ⁸ and determining that China had violated the Philippines' rights to an exclusive economic zone in parts of the South China Sea.Footnote ⁹ As these cases illustrate, international courts have rapidly evolved from sleepy bodies hearing the occasional low-level dispute to prominent international actors that increasingly render decisions of immense political, diplomatic, economic, and security importance.Footnote ¹⁰ This development, in turn, has attracted scholars from a wide variety of academic disciplines eager to conceptualize international courts' influence on international politics.

In describing and analyzing the Judicial Trilemma, we draw primarily upon insights from two research traditions that are typically neither juxtaposed nor applied to the study of international courts, rational design (RD) and judicial politics. To help orient the reader to the analysis that follows, we provide a very brief and necessarily selective introduction to these schools of thought.

Rational design builds directly upon rationalist and institutionalist writings within the field of international relations that understand international institutions as resulting from states' rational pursuit of their own self-interests. Early writings in this vein explored the conditions under which self-interested states would cooperate, and, drawing upon game theoretic insights, conceptualized international institutions as arrangements that render international cooperation more feasible and more durable.Footnote ¹¹ RD scholars extended this analysis, shifting the focus from identifying the conditions under which cooperation is possible to understanding the design of international institutions. The key analytic move in the rational design literature is to explain institutional design choices by reference to the underlying cooperation problems that states seek to solve.

To this end, early rational design writings linked several key features of institutional design, such as membership rules, the scope of issues covered, and the rules for controlling the institution, with the presence of one or more recurrent cooperation problems, such as distribution problems, enforcement problems, and information asymmetries.Footnote ¹² Much subsequent work in this tradition focuses on specific types of treaty clauses. Thus, for example, RD scholars have demonstrated that the presence of “flexibility clauses” in treaties, such as safeguard, exit, or reservation clauses, vary according to the relevance to the treaty of uncertainty about the future state of the world, such as the susceptibility of an issue area to new technological developments or scientific knowledge.Footnote ¹³

Among the treaty clauses examined by RD scholars, substantial attention has been devoted to exploring the presence (or absence) of dispute resolution clauses. In a series of pioneering studies, Barbara Koremenos has shown that the presence of dispute clauses varies systematically by issue area; specifically, nearly three-quarters of human rights treaties, roughly half of economic agreements, one-third of security treaties, and less than one-quarter of environmental treaties contain dispute clauses.Footnote ¹⁴ Koremenos also claims that the inclusion of dispute clauses responds in systematic ways to underlying enforcement, commitment, and uncertainty problems.Footnote ¹⁵

RD scholars have less frequently turned their attention to the design features of international courts. To the extent that scholars writing in this vein focus on tribunals, they typically examine individual elements of a court's institutional characteristics, often in isolation from other features. Thus, by way of example, a strand of writings identifies and describes a variety of ex ante and ex post mechanisms that states might use to influence judicial behavior, ranging from the careful delimitation of a court's jurisdictional reach to threats to ignore a court's judgment.Footnote ¹⁶ However, while specific individual features of international courts have been studied, substantially less attention has been devoted to explaining broad patterns of institutional design across tribunals.

We seek to build upon and extend this existing scholarship, in particular by emphasizing that the characteristics of international tribunals do not exist in isolation, but are interrelated. In this regard, two core points deserve emphasis. First, while the RD approach generally seeks to explain the presence or absence of individual design features, we argue that design features cannot be understood in isolation, because individual design features, such as renewable terms of office and open voting and dissent, can interact in important and sometimes unexpected ways. It is that interaction, more than the choice of any individual design feature in isolation, that interests us here. The second point concerns the identity of the actors who appropriately fall within the scope of inquiry. RD focuses exclusively on states as the rational designers of international institutions. To be sure, states design international courts, and we seek to understand and recreate their design choices along multiple dimensions. Understanding the founding moments of design, however, tells only part of the story, because the actual workings of international courts, and the interactions among those dimensions, depend in large part on the actions of the judges who subsequently hear disputes, interpret the law, and issue rulings. For insights on how to conceptualize and understand international judicial behavior, we look beyond the scholarship on institutional design, and draw inspiration from the rich interdisciplinary writings on judicial politics.

The judicial politics literature theorizes and empirically studies the relationships between judges and their political environments and interlocutors.Footnote ¹⁷ Greatly simplifying a diverse body of work, judicial politics scholars typically focus on explaining measurable judicial behavior. In doing so they theorize judges as utility-maximizing rational actors, seeking to realize their goals within the variable constraints imposed by institutions and constitutional structures. In scholarship on U.S. courts, in particular, there is a growing consensus on conceiving of judges as policy-seekers, interpreting the law in light of their ideology and policy preferences.Footnote ¹⁸ Judicial politics scholars differ, however, on the question whether judges are able to follow their sincere preferences freely or whether they are constrained in doing so by other actors, including the political (legislative and executive) branches of government. The “attitudinalist” school argues that U.S. federal judges, granted lifetime tenure, are free to vote and rule in line with their individual ideological or policy views, and that these attitudes are in fact the best predictors of judicial votes. By contrast, the “separation of powers” school argues that judges act strategically, and pursue policy goals within constraints imposed by the legislative and executive branches (which can, at least with respect to statutory interpretation, overturn unwelcome decisions) or broader public opinion (on which courts depend for legitimacy).Footnote ¹⁹

This ongoing debate is, to a significant extent, a dispute over the independence of the judiciary from the other branches of government. And indeed, judicial politics scholars have extensively theorized and empirically studied the question of judicial independence, broadly defined as the ability of judges to decide cases free of extralegal pressures from outside actors.Footnote ²⁰ They have, for example, theorized a number of motives that political actors might have for delegating substantial independence to judges, including the need to prevent a tyranny of the majority, or abuse of power by the executive or legislative branches, or to make legislative agreements lasting and credible.Footnote ²¹ Simply declaring judges to be independent, however, is likely to be insufficient insofar as one or more other actors have the ability and incentive to undermine that independence, and so scholars have identified institutional or constitutional protections, such as non-renewable terms of judicial office or fixed salaries, which reduce the ability of the political branches to manipulate the incentives of judges and undermine their independence.Footnote ²² There is a relatively broad consensus among judicial politics scholars that U.S. Supreme Court justices, who benefit from these and other protections, enjoy substantial independence from the executive and legislative branches, but there is greater controversy over the independence of judges in other systems, including those who serve on courts in the various U.S. states, many of whom are elected for fixed terms and subject to periodic reelection or retention elections, rendering them potentially subject to positive or negative sanctions for individual decisions.Footnote ²³ This question of renewable terms is closely linked to another core theme in the judicial politics literature, that of judicial accountability (either to political branches of government or to some other stakeholders), which is frequently depicted as being in tension with (or indeed the flip side of) judicial independence.Footnote ²⁴ These twin themes, of judicial independence and judicial accountability, and the potential tradeoff between them, are central to our argument in this article, and we shall return to them below.

The judicial politics literature originated in the study of American courts, yet has recently expanded to encompass the comparative study of constitutional courts in advanced democracies,Footnote ²⁵ and more recently to the study of international courts.Footnote ²⁶ Judicial politics students of international courts have, for example, studied the processes of judicial appointment to various international courts,Footnote ²⁷ the use of precedent by international judges,Footnote ²⁸ and the independence and accountability of the Court of Justice of the European Union.Footnote ²⁹

For our purposes, the judicial politics literature is doubly valuable, both for situating the judiciary in its political context, and for its careful analyses of the independence and accountability of judges vis-à-vis political branches of government. Yet, in large part because of its roots in the study of American federal courts, the judicial politics literature has paid far less attention to several other important questions, including the impacts of limited and renewable terms of office; the significance of open voting and separate opinions, or what we shall call judicial transparency, which are universal in American courts but an important variable in foreign and international courts; and above all the complex interactions among judicial independence, judicial accountability, and judicial transparency. As we shall see, it is precisely these interactions that give rise to the Judicial Trilemma.

B. Tribunal Characteristics Relevant to the Judicial Trilemma

The Judicial Trilemma results from the interplay among three specific values that states might seek to embed in any international tribunal, and three specific associated characteristics that any court might possess. The first is judicial independence. In principle, we assume that when states create international courts, they seek to populate them with independent judges to conduct unbiased third-party dispute settlement.Footnote ³⁰ They do so because states recognize that in an anarchic international setting lacking centralized oversight and enforcement mechanisms, independent courts permit states to enhance the credibility of the commitments they make to other states.Footnote ³¹ Since independent courts can detect, identify, and publicize violations, they tend to increase the cost of violations and therefore increase the probability that states will comply with their obligations. By incentivizing compliance with obligations, independent courts can increase the value of international cooperation.Footnote ³²

Of course, to say that states value judicial independence is emphatically not to say that international judges should be entirely unconstrained. For example, it is not acceptable for judges to act in partisan or self-interested ways. Thus, while judicial independence is central to virtually all plausible theories of well-functioning courts, it is equally fundamental that robust checks on the exercise of power, including judicial power, are necessary.Footnote ³³ For this reason, the second feature of interest is judicial accountability. In this context, accountability means that “some actors have the right to hold other actors to a set of standards, to judge whether they have fulfilled their responsibilities in light of those standards, and to impose sanctions if they determine that these responsibilities have not been met.”Footnote ³⁴

The literature identifies numerous ways that courts as institutions can be held accountable to the states that create and use them. Tribunal accountability can come in many forms, including “financial accountability (to the budgetary authority), case-management accountability (for the efficient processing of cases), … process accountability (for procedures), and ‘content’ accountability (for individual judgments),” and many mechanisms can be used to promote these various forms of accountability.Footnote ³⁵

Judicial accountability, in contrast, focuses on the individual judge, not the court as an institution. It is thus a substantially narrower category and primarily involves judicial terms of office and the possibility of reappointment. We foreground judicial reappointment processes for several reasons. First, unlike U.S federal judges who enjoy lifetime appointments, judges at virtually every international court serve for limited time periods, most with the possibility of renewal at the end of a term. Moreover, judges at most international courts cannot be removed during their terms of office, except by the unanimous vote of their fellow judges, and only when they are “unable to fulfill the required conditions,” such as being unable to perform the duties of the office. Thus, reappointment is the primary mechanism whereby member states can hold individual judges accountable for their behavior.Footnote ³⁶ Second, judicial appointment provisions, including both length and renewability of terms, vary considerably across tribunals. For example, WTO AB members serve for four-year terms, and may be reappointed once, while CJEU judges serve six-year terms, and are eligible for reappointment. ICJ and ITLOS judges serve for nine-year terms, and may be reappointed, while ECtHR and International Criminal Court (ICC) judges also serve for nine-year terms, but are not eligible for reappointment.

For purposes of explicating the Trilemma, we focus on whether judicial terms are renewable as a threshold and defining dimension of judicial accountability.Footnote ³⁷ This conceptualization is not, however, intended to suggest that judicial accountability can or should only be thought of as a dichotomous variable. To the contrary, variation in institutional design features at different courts can determine how often, and to whom, judges are accountable. By way of example, we highlight two such features, the length of judicial terms and the rules governing renomination and reelection to the bench.

First, and most obviously, different term lengths create different degrees of judicial accountability. Among international courts that permit reelection, states have provided for substantially different lengths of term. Assuming that the desire to remain in office might influence judicial behavior, institutional variation in term lengths creates different incentives for judicial behavior on different courts, and correspondingly different levels of judicial accountability to member states.

Second, the court-specific provisions for judicial appointment and reappointment determine to whom judges are accountable. As a general rule, appointment to an international court is a two-stage process, in which one set of actors (typically individual member governments) first nominate judges, and then a broader electorate or selectorate (typically a plenary assembly of the member states) makes the final selection. Within this general framework, however, the statutes of different international courts vary substantially, which in turn determines which actors have the ability to approve, or veto, the reappointment of individual judges.

The bench of the CJEU, for example, consists of one judge from each member state. A potential candidate for the CJEU must obtain the support of his or her home government to be nominated (or renominated). Once the government forwards an individual as a candidate for the bench, however, the long-established practice is for other states to acquiesce in this choice. Thus, the nominee (and potential renominee) is de facto accountable to only one state, which is effectively responsible for deciding whether the individual will be nominated or renominated. In this case, we might say that there is strong and direct accountability to the home state, but only very limited accountability to other states.

At the ICJ, in contrast, only a fraction of states will have a judge of their own nationality on the bench, and the election and reelection process has a different dynamic. As a practical matter, an individual must obtain the support of her home government to be nominated or renominated.Footnote ³⁸ Typically, more individuals are nominated than can serve, and to be elected or reelected, a candidate must receive a majority of the votes of the United Nations General Assembly and the Security Council. Hence, a judge with an eye to renomination would consider the preferences of his or her own state, but would in addition be cognizant of the preferences of the broader UN membership. In this case, we might say there is strong and direct accountability to the judge's nominating state, and diffuse accountability to the electorate as a whole.

At the WTO, an AB member needs the support of his/her own state at the nomination stage; once nominated, however, the WTO's consensus rule means that a single WTO member state can block a judge's appointment. Much the same process is followed for reappointments, including the ability of any state to block a reappointment under the WTO's consensus rule. In this context, therefore, there is strong and direct accountability to the home state, and strong accountability to each and every one of the voting states.

Hence, even within the category of high accountability courts, different structural features, prominently including length of term and voting rules for election to the bench, can interact to generate a continuum of judicial accountability. In graphic form, this variation can be presented as follows:

Table 1.

Variation in Judicial Accountability at Selected International Courts

The third feature of interest is judicial transparency. The term “transparency” has many meanings in international legal discourse,Footnote ³⁹ and even in the context of international adjudication, transparency can refer to who can submit written pleadings, whether oral proceedings are broadcast or otherwise accessible to the public, the nature of the judges' deliberative processes, and the means used to disseminate the court's judgments, among other topics. In highlighting judicial transparency, we focus on a more limited and precise aspect of transparency, namely the ability to identify a particular judge's position or vote on a particular issue before the court, which might be called judicial identifiability. The most visible and common mechanisms related to identifiability are found in the format and content of a court's judgments, in particular the public reporting of judicial voting and the use of separate concurring or dissenting opinions.

Here again, international courts display enormous variation. At some courts, such as the ICJ and the International Tribunal for the Law of the Sea, judgments explicitly state the number and the names of the judges constituting the majority and the minority on each operative provision of the judgment. Moreover, virtually every judgment at both of these courts is accompanied by one or more separate opinions. In these instances, every judge's position on virtually every important issue in a case is publicized, and judicial identifiability is extremely high. In stark contrast, at other courts, such as the CJEU, judgments are always issued in the name of the court, no indication of the judges' votes on any issue appears, and separate concurring or dissenting opinions are never issued. In these courts, judicial identifiability is extremely low. Still other courts occupy intermediate positions. For example, ECtHR rules require that a judgment state “the number of judges constituting the majority,” but not their names, and both the court's Statute and the judge-made Rules of Court explicitly allow for separate concurring or dissenting opinions, which are commonplace. At the WTO's AB, identifiability is lower, as separate opinions are issued in fewer than 10 percent of the reports issued by the Appellate Body and, when issued, are anonymous.

In practice, the two aspects of judicial transparency—open voting and the possibility of issuing separate opinions—although in principle separable, tend to co-vary among the courts we studied, such that high-transparency courts (such as the ICJ) feature both open voting and the possibility of writing separately, while low-transparency courts neither publicize judicial votes nor issue separate opinions with their judgments. We focus in this article primarily on the issue of separate opinions, which reveal the legal reasoning as well as the votes of judges who disagree with a majority opinion.

A large debate surrounds the normative desirability of judicial transparency, and in particular the desirability of separate concurring or dissenting opinions.Footnote ⁴⁰ Advocates of dissent claim that they improve the quality of a court's reasoning, as well-reasoned dissents force the majority to grapple with the strongest arguments posed by the losing side.Footnote ⁴¹ Moreover, dissents can provide practical guidance to litigants and other courts, identifying potential lines of argument to pursue in future cases. Finally, dissents can influence the development of legal doctrine. The most enduring dissents serve as “appeal[s] to the brooding spirit of the law, to the intelligence of a future day,” when the arguments presented in a dissent may one day persuade a future majority, and become the law.Footnote ⁴²

Against this catalogue of asserted advantages, other judges and scholars offer a list of objections to dissent. First, a fractured court can damage a court's legitimacy, particularly early in a court's history, before it has created a legacy; “disunity cancels the impact of monolithic solidarity on which the authority of a bench of judges so largely depends.”Footnote ⁴³ Moreover, dissents can sow confusion among bench and bar, as a splintered set of opinions can leave the precise state of the law unclear. Dissents can also undermine judicial collegiality, as judges rarely appreciate having their errors pointed out in a highly public fashion, and frequent dissents can contribute to divisiveness and ill feelings on a court.

More importantly for current purposes, judicial identifiability also provides a pathway into understanding the logic that drives the Judicial Trilemma. For judges who enjoy life tenure, issuing individually signed, public, and potentially unpopular opinions—whether majority, concurring, or dissenting—is unlikely to pose any significant threat to judicial independence, since the judge is insulated from retaliation or reward by the political branches or citizenry. By contrast, individually signed public opinions may substantially threaten the reappointment prospects of judges who lack life tenure. For example, state courts in the United States provide many examples of judges who were not reelected following the issuance of unpopular opinions.Footnote ⁴⁴

Many judicial systems have struggled to strike the appropriate balance between permitting the issuance of separate, potentially unpopular opinions, on the one hand, and the nature and structure of judicial terms and appointment procedures, on the other. For example, for many years, judges at the German Constitutional Court were statutorily prohibited from issuing separate opinions. In 1970, the German legislature enacted a reform that allowed the Court to publish concurring and dissenting opinions. At the same time, “[r]esponding to the concern that reappointment considerations might influence a judge's votes … ,”Footnote ⁴⁵ the political branches altered the appointment process for those judges, moving from a system of short, renewable terms to a new system of non-renewable, twelve-year terms. The same dynamic is seen at international courts, where states may retaliate against decisions they do not like by opposing a judge's reappointment—as the case of Seung Wha Chang vividly illustrates.

C. The Logic of the Judicial Trilemma

Building on the three core concepts above, and drawing inductively on interviews with current and former international court judges and court officials, we may theorize more generally about the interrelationship between judicial appointment, judicial independence, and judicial transparency. More specifically, we hypothesize that with respect to these three characteristics, judicial systems face potential tradeoffs, such that any given court can maximize two, but not all three, of these features (see Figure 1).

Figure 1.

The Judicial Trilemma: “Pick two, any two”

In principle, we can imagine that states designing international courts might wish to create courts that exhibit high levels of judicial accountability, judicial transparency, and judicial independence. But the logic of the Trilemma suggests that it is not possible to maximize all three of these values at once. By way of example, imagine that states design a court with high judicial accountability, by making the judges subject to periodic assessment and reappointment by the members after a fixed term of office, and with high judicial transparency, by requiring publication of the judges' individual votes and by compelling or allowing the publication of individual dissenting or concurring opinions. By maximizing these two features, however, the member states render the judges vulnerable to individual assessment by member states, which may then opt to block the reappointment of individual judges in response to unwelcome rulings. Put simply, assuming that judges are at least partially motivated by a desire to retain their positions (an assumption to which we return below), efforts to maximize both judicial accountability and judicial transparency inherently place pressures on judicial independence. If this logic is correct—as many of the judges we have interviewed believe—then states that design international courts, and the judges that serve on those courts, can choose to maximize two important values, but only at the cost of some sacrifice to a third important value. And if this characterization of the strategic environment in which states and judges find themselves is accurate, then most international courts (and indeed domestic courts) can approximate one of three ideal types:

First, if states most highly value independence and accountability, they can design a court whose judges have limited, renewable terms of office (hence maximizing accountability), and can protect those judges' independence by minimizing judicial transparency or identifiability. In these cases, states can design court statutes that either compel, or at least allow, judges to issue per curiam decisions, such that neither the vote nor the opinions of individual judges can be identified. For their part, judges with limited, renewable terms of office can choose to reduce their vulnerability to retaliation for unwelcome rulings by reducing (insofar as they are able) the transparency of their court's decision-making, by issuing per curiam rulings and suppressing any sign of individual votes or positions, even where their statutes explicitly allow for open voting and/or separate opinions. This is the choice made by the designers, or the judges, of traditional, Continental European civil law courts (including the pre-1970 German Constitutional Court), all of which opted deliberately to reduce judicial identifiability by suppressing public dissent. Among international courts, as we shall see, the CJEU exemplifies this ideal type, suppressing for over seven decades the individual votes and views of the judges and issuing per curiam rulings in an effort to protect the independence of judges with renewable six-year terms.

Second, if states most highly value independence and transparency, they may design courts whose statutes compel or allow judges to vote openly and to issue individual concurring or dissenting opinions, thereby allowing the judges to be individually identified. In these cases, states that value an independent judiciary will be forced to sacrifice judicial accountability, for the simple reason that individually identifiable judges with renewable terms will face incentives to issue votes and rulings that conform to the interests of the states responsible for renomination and reelection. In these cases, we should expect judges to have non-renewable terms, which may be either for life or for a fixed term of office. Among domestic courts, the United States Supreme Court, whose judges enjoy life tenure (except for impeachable offenses) and the post-1970 German Constitutional Court, whose judges have non-renewable twelve-year terms, both fit this model, voting and dissenting openly and without fear of retribution due to their non-renewable terms. Among international courts, both the post-2010 European Court of Human Rights, which we will examine in detail below, and the International Criminal Court fit this pattern. Judges at both courts serve for non-renewable nine-year terms, and frequently issue judgments that are accompanied by one or more dissenting or concurring opinions.

Third and finally, states may choose to maximize both judicial accountability and judicial transparency, by designing courts whose judges have both renewable terms of office and a statute featuring the requirement or possibility of open voting and separate opinions. The choice to maximize accountability and transparency comes at a potential cost, however, leaving judges vulnerable to member state retaliation for identifiable, individual votes or opinions. In domestic jurisdictions, one approximation of this ideal type is found in the courts of those individual U.S. states where judges vote openly and write separately (hence, high transparency) and serve fixed, renewable terms of office before facing reelection or reappointment by executives, legislatures, or electorates (hence, high accountability). While this design choice has its defenders, critics point out that this combination of design features imperils the independence of judges, by placing them under pressure to offer judicial rulings that will be acceptable to whichever electorate or selectorate is entrusted with their reappointment.Footnote ⁴⁶ This ideal type also comes close to identifying an important and large group of international courts, whose judges are required or allowed to vote and write individually and openly, but who are subject to renomination (typically by their home governments) and reelection (typically by a general assembly of the members). To be clear, we do not mean to suggest that judges on these courts, which include the ICJ and ITLOS among others, should be understood as obedient servants of their political masters, but rather that the design of their courts to maximize judicial accountability and transparency creates a structural context that poses certain risks to judicial independence by exposing individual judges to retaliation for unwelcome rulings.

In tabular form, the logic of the Trilemma suggests three ideal-type courts, each of which achieves high values for two of the three features, at the expense of low values for the third (see Table 2).

Table 2.

The Judicial Trilemma: Three Ideal-Type Courts, with Examples

As formulated here, the Judicial Trilemma is a heuristic device that illuminates both the design choices available to states creating international courts and the choices of international judges seeking to carry out their mandates within the limits of their respective statutes. We ground this heuristic both in the logical argument spelled out above, and in the testimony and lived experiences of the international court judges who have confided in us about the tensions they encounter in their day-to-day judicial practices among judicial accountability, judicial transparency, and judicial independence.Footnote ⁴⁷ Significantly, unlike the trilemma models found in economic theory,Footnote ⁴⁸ our Judicial Trilemma heuristic does not rely on highly restrictive assumptions about the preferences of either states or judges. Instead, we argue, the basic logic of the Trilemma is consistent with minimal assumptions about the motivations of both judges and states.

With respect to judges, our analysis does not rely on narrow or controversial assumptions about judges being either pure policy-seekers or Posnerian careerists.Footnote ⁴⁹ Indeed, we believe that international judges have quite heterogeneous motivations, and that even for any specific judge, preferences may change over the course of a career, or in response to changing circumstances. Hence, we are largely agnostic about judicial preferences, and our analysis rests on the very minimal assumption that, whatever their intrinsic motivations (ideology, policy-seeking, reputation, remuneration, etc.), judges value continuing in their positions as a necessary condition to achieving many, if not all, of these goals. Hence, while individual judges will vary in their specific preferences, we join the majority of judicial politics scholars in assuming that the median judge is likely to be influenced, at least to some degree, by the prospect of reappointment at the end of a fixed, renewable term of office.Footnote ⁵⁰

Turning from judges to states, we similarly make no strong assumptions about the preferences of states for any of our three values, or about states operating with perfect information when designing courts or appointing judges. Indeed, individual states may well have different preferences as among judicial independence, accountability and transparency, as a function of their regime type, legal tradition, etc.Footnote ⁵¹ Our argument is simply that states need to choose as among our three values, not that they will always and everywhere make a single, predictable choice.Footnote ⁵²

Finally, our analysis does not posit or require normative commensurability among the three values; that is, we do not wish to claim that the three values are of equal importance to us, to states, or to judges. Nor would we claim that judicial accountability, transparency and independence are the only values that international courts might embody. Once again, our central claim is simply that it is not possible simultaneously to maximize these three values, that they coexist in an uneasy tension, and that states and judges necessarily need to make choices among them.

A. The Court of Justice of the European Union

The Court of Justice of the European Union (formerly called the European Court of Justice (ECJ)) is the judicial organ of the European Union.Footnote ⁵³ Its primary task is to examine the legality of European Union (EU) measures and to ensure the uniform interpretation and application of EU law. The Court consists of twenty-eight Judges, one from each EU member state. It is assisted by eleven Advocates General, who present the Court with impartial and independent “opinions” in the cases assigned to them.

The Court was created, and its original statute was drafted, in the early 1950s, as part of the 1951 Treaty of Paris which created the European Coal and Steel Community (ECSC) and its first supranational High Authority. The founders of this first Community envisaged a court whose primary role would be administrative, ruling on the legality of the decisions of the High Authority, rather than interstate dispute settlement. Accordingly, they designed the first Court's statute on the model of the French Conseil d’État, and many of the member states' original design choices—which have for the most part carried forward into the much larger and busier Court of today—bear the imprint of the French civil law model.Footnote ⁵⁴

The CJEU has been enormously influential. Numerous scholars have detailed the Court's leading role in the gradual “transformation” of Europe, including through the Court's development of the doctrines of direct effect and supremacy.Footnote ⁵⁵ Political scientists have detailed how the Court's decisions helped push European political and economic integration further and faster than member states had been prepared to go on their own.Footnote ⁵⁶ A broad scholarly consensus concludes that

The CJEU exhibits one of the three “ideal” type responses to the Trilemma, combining high levels of judicial independence and judicial accountability with low levels of judicial transparency. With respect to judicial independence, the drafters of the Paris Treaty, and those of the Court's subsequent statutes which have largely hewn to those original provisions, adopted a text that emphasized and sought institutional protections for judges' individual and collective independence. Thus, Article 19 of today's Treaty on the European Union and Article 253 of the Treaty on the Functioning of the European Union provide that CJEU judges “shall be chosen from persons whose independence is beyond doubt,” and this theme is reinforced in the Court's Statute, which provides that before taking up his duties, a judge shall, in open court, take an oath to perform his duties impartially and conscientiously, and which prohibits a judge from exercising any political or administrative function, or from engaging in any occupation (absent a specific exemption granted by the Council). The judges themselves have evidenced their commitment to judicial independence by adopting a Code of Conduct which includes several obligations with a view to guaranteeing independence and impartiality, including the filing of a declaration of financial interest, and limitations on the type of activities that current and former judges can engage in.

Other provisions enhance this commitment to judicial independence. For example, CJEU judges enjoy immunity from legal proceedings while in office, and after leaving office they “continue to enjoy immunity in respect of acts performed by them in their official capacity, including words spoken or written.”Footnote ⁵⁸ Moreover, judges enjoy protections from removal during their terms of office. A judge may be removed “only if, in the unanimous opinion of the Judges and Advocates-General of the Court of Justice, he no longer fulfills the requisite conditions or meets the obligations arising from his office.”Footnote ⁵⁹

In terms of judicial accountability, the Court's Statute provides that judges serve for relatively short six-year terms, which are renewable.Footnote ⁶⁰ Each government has the de facto ability to appoint a national judge to the court. Although appointments are formally by common accord of all governments, traditionally national candidates are rarely, if ever, second-guessed by other states.

The Lisbon Treaty modified this system somewhat by introducing the so-called Article 255 panel procedure. Under this procedure, operational since 2010, a seven-member panel, consisting of former national supreme and constitutional court judges and former ECJ judges, interviews candidates for appointment or reappointment and delivers a reasoned opinion on the candidate's suitability. Several candidates have received negative opinions, and these individuals were replaced by their national governments.Footnote ⁶¹ While the panel provides a “quality check” on the process, governments still retain the ability to nominate and renominate national judges, including the unfettered discretion not to reappoint a judge should the government disagree with that judge's rulings in particular cases.

As the Trilemma would suggest, given the Court's high independence/high accountability structure, the CJEU is a “low transparency” court. The Court's Statute provides that judgments “shall contain the names of the Judges who took part in the deliberations.” Unlike the statutes of some other international courts, however, it does not provide that the judgment shall identify who was in the majority on any particular issue. Just as importantly, the Court's original 1951 Statute, and all of those that have followed, make no reference at all to the possibility of judges issuing separate concurring or dissenting opinions. To some extent, this omission can be attributed to the Court's origins as an administrative court modeled on the French Conseil d’État: the common-law tradition of separate opinions was alien to the French and other Continental European legal systems of the original six member states in the early 1950s, and in this sense it is unsurprising that the original member states made no mention of separate opinions in the CJEU Statute.Footnote ⁶²

All CJEU judgments, from its inception to the present day, have been issued in the name of the Court, and not a single judgment has included a separate concurring or dissenting opinion. Some CJEU judges we interviewed claimed that the Court's Statute precludes the writing of separate opinions, citing a provision stating that “[t]he deliberations of the Court of Justice shall be and shall remain secret.”Footnote ⁶³ These judges argue that publication of separate opinions would necessarily reveal, at least to some extent, elements of the court's deliberations, in particular the identity and arguments of those who depart from the majority of the court on specific issues.Footnote ⁶⁴ While this is surely a plausible reading of the Statute, the relevant language hardly compels this conclusion. In fact, the statutes and rules of other international courts contain similar language regarding the secrecy of deliberations,Footnote ⁶⁵ yet judges on these courts routinely issue separate opinions without triggering claims that doing so violates the secrecy of deliberations.

Our interviews with CJEU judges, reinforced by a large body of scholarly writings,Footnote ⁶⁶ suggest a different motivation. Specifically, the judges recognized the logic of the Trilemma and, given their desire for independence and their relatively short terms of office, decided on their own not to issue separate decisions. Indeed, many former judges have candidly suggested as much in writings that address the interplay of short, renewable terms, the issuing of per curiam rulings, and the fragility of judicial independence.Footnote ⁶⁷ Consider, for example, Judge Azizi, who suggests that “the very fact that judges may be reappointed could be seen as putting at risk their independence,” and that

Substantially similar sentiments were expressed by nearly all of the CJEU judges we interviewed. Several suggested that, if dissents were allowed, judges might feel pressure to write separately on cases of importance to their home states, particularly as they approach the end of their terms.Footnote ⁶⁹ Avoiding such a conflict was, accordingly, the main, although not the only, reason the judges cited for avoiding public dissent.Footnote ⁷⁰

The argument that renewable terms represent both the primary threat to judicial independence and the primary impediment to the introduction of dissents appears in much of the scholarship on the Court as well. Writing in 2001, for example, Joseph Weiler argued for the introduction of dissenting opinions to improve the clarity of the Court's famously brief and sometimes cryptic legal reasoning,Footnote ⁷¹ but noted that a “precondition” to any such reform would be the elimination of renewable judicial terms, which he called “a continuous affront to the integrity of the European legal system.”Footnote ⁷² Writing more than a decade later, Weiler repeated his call for the introduction of separate and dissenting opinions, but here again concluded that, “So long … as the judges may be reappointed the possibility of dissenting opinions would be inimical.”Footnote ⁷³

Thus, the CJEU—both in its original design, and in the subsequent behavior of the judges—represents one response to the Trilemma, associated with high accountability, high independence, and low transparency. The original 1951 Statute of the Court, unchanged in its fundamentals today, created a Court with high accountability of judges to their home member states, and an aspiration of high independence, combined with language that allowed for low transparency, given the absence of open voting and the silence on the possibility of separate opinions. In the subsequent decades, moreover, the judges, profoundly aware of the vulnerability attached to six-year renewable terms, took a deliberate decision to pursue a “low transparency” strategy of suppressing all records of judicial votes, and all signs of internal dissent on the bench, issuing all judgments in the name of the Court. Reappointing states, in turn, lack the ability to “retaliate” against a national judge due to his or her positions, and judicial independence is maintained. Graphically, this approach can be represented as follows:

Figure 2.

The Judicial Trilemma at the Court of Justice of the European Union

Significantly, both judges and scholars have argued for a change to the judges' six-year renewable terms in favor of longer, non-renewable terms of nine or (preferably) twelve years that would reduce their accountability to their home states.Footnote ⁷⁴ In its report to the 1996 Intergovernmental Conference that culminated in the Treaty of Amsterdam, for example, the Court expressed an interest in moving toward longer, non-renewable judicial terms of office, a reform that it justified explicitly in terms of judicial independence.Footnote ⁷⁵ The European Parliament, several member governments, and a number of nonstate actors offered similar proposals,Footnote ⁷⁶ but the member states made no changes in this regard—a failure for which they have been excoriated by Weiler.Footnote ⁷⁷ However we normatively interpret member states' refusal to move toward non-renewable terms of office, it is clear that in the absence of such a change, the de facto ban on dissents represents a robust equilibrium, broadly supported by judges and scholars alike. As we shall see, however, other international courts have adopted different approaches.

B. The European Court of Human Rights

The European Court of Human Rights hears cases alleging violations of the European Convention on Human Rights (ECHR), a treaty whereby the (now) forty-seven member states of the Council of Europe are obliged to uphold fundamental civil and political rights of individuals under their jurisdiction. Since the Court began operations in 1959, the Convention has been amended several times, and its supervisory and enforcement mechanisms have likewise undergone a considerable evolution. Below, we discuss some of the key moments in this evolution, and explore their implications for the Trilemma.

During negotiations over the original Convention in 1949–1950, the initial members of the Council of Europe were deeply divided, with certain “maximalist” members seeking the creation of a strong human rights court with compulsory jurisdiction and a right of individual initiative, while other states opposed the creation of any court at all. In the compromise that followed, a Court was created, but both compulsory jurisdiction and individual access to the Court were made the subject of optional protocols. Furthermore, access to the Court would be mediated by a quasi-judicial European Commission on Human Rights, which would screen cases and, if its members considered a case to be well-founded, would forward that case to the Court on the individual's behalf. From the 1950s to the early 1970s, few governments accepted the ECtHR's compulsory jurisdiction, the Commission dealt with most cases on its own, leaving the Court substantially underutilized.

Over the next decade, the pace of activity increased, and by the late 1970s the Court had found violations of the Convention in a number of high-profile cases. Moreover, as more states accepted the ECtHR's compulsory jurisdiction, and the number of cases increased, the system began to clog, and the two-tier process was seen as overburdened and unacceptably slow. Substantial debate led a majority of European states to accept Protocol 11 to the European Convention, which eliminated the Commission, required all member states to accept the ECtHR's compulsory jurisdiction, authorized individuals to file cases directly with the Court, and transformed the ECtHR into a full-time body.Footnote ⁷⁸

Russia and a number of Central and Eastern European states ratified the Convention in the 1990s, extending the Court's reach far beyond Western Europe. The combination of an expanded number of member states and the right of individual petition, however, generated an enormous backlog of cases, and by 2000, there was general agreement that the Court confronted “a major crisis.” As a result, in May 2004 the Council of Europe Committee of Ministers adopted Protocol 14 to the European Convention.Footnote ⁷⁹ Its aim was to improve the Court's efficiency, in part by strengthening the Court's ability to quickly dispose of applications that are clearly inadmissible.

Over nearly six decades, the ECtHR has been widely acclaimed, and often characterized as “the world's most effective international human rights tribunal”Footnote ⁸⁰ and as a quasi-constitutional court for Europe. During that period, the Court has taken an expansive (some would say activist) approach to the treaty, gradually expanding the meaning of the core rights contained in the Convention's list of fundamental human rights.Footnote ⁸¹

For current purposes, the ECtHR represents a second approach to the Judicial Trilemma, combining high levels of judicial independence and judicial transparency with levels of judicial accountability that have varied over time (see below), but are today far lower than those at the CJEU. Let us consider each, briefly, in turn.

With respect to judicial independence, the European Convention does not explicitly mention judicial independence, providing only that judges “shall be of high moral character and must either possess the qualifications required for appointment to high judicial office or be jurisconsults of recognized competence.”Footnote ⁸² Upon taking office in 1959, the judges adopted their own Rules of Court, which require each judge, before taking his duties, to take an oath or solemn declaration to “exercise [her] functions as a judge honourably, independently and impartially,” language that remains essentially unchanged today.Footnote ⁸³ Later rules provided that judges shall serve in their individual capacity and that they should not “engage in any political or administrative activity or any professional activity which is incompatible with their independence or impartiality. …”Footnote ⁸⁴ Protocol 11 strengthened the commitment to judicial independence by providing that judges shall sit in their individual capacity, and “shall not engage in any activity which is incompatible with their independence, impartiality or with the demands of a fulltime office.”Footnote ⁸⁵

A number of other structural features are intended to buttress judicial independence. First, ECtHR judges enjoy privileges and immunities in the exercise of their official functions. They enjoy immunity from legal process for official statements, and are exempt from taxation and immigration restrictions. Even after they leave office, they continue to enjoy immunity for “words spoken or written and all acts done by them in discharging their duties.” Second, they enjoy protections from removal from office. The Convention provides that no judge may be dismissed from office unless a supermajority of two-thirds of the other judges decide that the judge has ceased to fulfill the required conditions. To date, no judge has been dismissed under this provision.

In fact, from the Court's earliest days, there was little doubt regarding judicial independence. A bold series of early decisions against member states firmly established that judges were willing to rule against member governments, and by the 1970s commentators were already declaring that the Court had established “a reputation for quality and independence.”Footnote ⁸⁶

The ECtHR is also a high transparency court. The original Convention said little about the form of judgments. Specifically, it did not direct that judgments list the judges who participated, or how they voted. However, the Court's Statute, modeled explicitly on the Statute of the ICJ,Footnote ⁸⁷ did provide that “[i]f a judgment does not represent in whole or in part the unanimous opinion of the judges, any judge shall be entitled to deliver a separate opinion.”Footnote ⁸⁸

The initial Rules of Court, adopted in 1959, tilted further in the direction of transparency by providing that judgments should include the names of the judges constituting the Chamber and the number of judges constituting the majority, thus revealing whether a decision had been unanimous, although not the names of the judges in the majority or the minority. The rules also allowed for judges to issue separate concurring or dissenting opinions or a “bare statement of dissent.” This language, which remains essentially unchanged in Article 74 of the current Rules of Court, did not explicitly require judges in the minority to publicly declare themselves as dissenters (although some current judges interpret the rule in precisely that way), but neither did it discourage dissent.

From the earliest days, ECtHR decisions were commonly accompanied by separate dissenting and concurring opinions. Between 1959 and April 2001, the Court issued just over 2,000 judgments, of which 602, or nearly one third, had one or more separate, dissenting, or concurring opinions.Footnote ⁸⁹ Interestingly, these patterns did not significantly change in the aftermath of Protocol 11. Of the cases decided between November 1, 1998 and October 31 2001, 70 percent were unanimous, and 30 percent had dissenting or concurring opinions.Footnote ⁹⁰ Interviews with judges on the contemporary Court similarly reveal a highly permissive attitude toward dissent, with many judges indicating to us that they feel no reticence about issuing dissenting or concurring opinions, subject only to informal norms about the length and respectful tenor of such opinions, or indeed that they feel obligated, if they vote against the majority, to issue at least a brief separate opinion explaining why.Footnote ⁹¹

Given both high judicial independence and high judicial transparency, the logic of the Judicial Trilemma would predict low judicial accountability. However, the states that negotiated the ECHR in 1950 created a high-accountability court. The original Convention provided that ECtHR judges would serve for renewable nine-year terms.Footnote ⁹² In Protocol 11, moreover, the member states opted to increase the accountability of the judges, by changing the term of office to a six-year renewable term.Footnote ⁹³ It is understandable that at the very time that states were substantially strengthening the Court, they would seek a corresponding increase in judicial accountability, yet the change was criticized at the time for potentially diminishing the independence of judges vis-à-vis their home states.Footnote ⁹⁴

Thus, the early experience of the ECtHR seems to confound the Trilemma, as member states chose, and the Court seemed to enjoy, high levels of independence, accountability, and transparency. But this constellation of features was not sustainable. Specifically, by the late 1990s and early 2000s, Court-watchers began to voice concerns that ECHR member states were using the reappointment power to retaliate against national judges whose opinions they found objectionable. As Erik Voeten reported:

Faced with evidence of national governments seemingly punishing judges for unwelcome decisions, the Court's allies grew alarmed, and in 2001 the three-member Evaluation Group, appointed by the Council of Europe to study the Court's workings and recommend amendments to the Convention, proposed a change to the terms of the judges from a renewable six-year term to a non-renewable nine-year term, justifying the change explicitly in the interest of ensuring the independence of the judges:

This proposal was championed in turn by the Council of Europe's Parliamentary Assembly,Footnote ⁹⁷ which advocated nine-year non-renewable terms for the judges to be able to act “without fear or favour,” and finally endorsed by the ECHR member states, which included the reform as part of Protocol 14 in 2004. In place of six-year, renewable terms, the Protocol provides for a single non-renewable term of nine years. The explicit rationale for this change—the first such reform in the history of the international judiciary—was a concern about judicial independence.Footnote ⁹⁸ As one set of commentators diplomatically put it, the change from renewable to non-renewable terms “ha[d] its origins in concerns at the Court, the Parliamentary Assembly and the Committee of Ministers about a few instances where there seemed to be abuse of the current provisions in that sitting judges of recognized competence and effectiveness had not been included on the list of candidates for judge for their country on expiry of their term, apparently for purely political reasons.”Footnote ⁹⁹ Others have more straightforwardly noted that the change was intended to “offer more guarantees for the Court's independence,”Footnote ¹⁰⁰ and “to reinforce the independence and impartiality of the judges [as] the situation in which judges had to resubmit their candidatures to the government was perceived as awkward and undignified.”Footnote ¹⁰¹

Thus, the post-Protocol 14 ECtHR represents a second “ideal” type response to the Trilemma. Initial design choices in the 1950s, which combined high levels of accountability and judicial transparency, eventually produced an unintended but significant threat to a third value, judicial independence. Judges, parliamentarians, and a large majority of member governments, pushed back against this pressure. In response, states agreed to change judicial terms of office, making them longer and non-renewable. These changes, in turn, have alleviated, although not entirely eliminated, political pressures on judicial independence.Footnote ¹⁰² Graphically, we can represent the ECtHR's contemporary (post-2010) response to the Trilemma as in Figure 3:

Figure 3.

The Judicial Trilemma at the European Court of Human Rights

Comparing the ECJ and ECtHR cases, we see that both member states and judges of each Court came to recognize, more or less explicitly, the stark tradeoffs of the judicial trilemma, although the two courts responded to this trilemma in strikingly different ways. At the Luxembourg court, judges appointed for short, renewable terms have taken a conscious decision to remain (as they have always been) a low-transparency court, interpreting the secrecy of deliberations to prohibit any public sign of dissent or disagreement within the court, and in the process protecting individual judges from retaliation for unwelcome rulings. In Strasbourg, by contrast, the spate of non-renewals after 1998 forced member states as well as judges to confront the nature of the trilemma, to which they responded with an unprecedented decision to shift to non-renewable judicial terms of office, effectively accepting less judicial accountability in return for high judicial transparency and independence. We now examine a court that represents a third response to the Trilemma.

C. The International Court of Justice

The International Court of Justice is “the principal judicial organ of the United Nations.”Footnote ¹⁰³ The Court's Statute draws heavily upon the Statute of the Permanent Court of International Justice (PCIJ),Footnote ¹⁰⁴ a tribunal that was associated with the League of Nations and served as the first standing international court with general jurisdiction. The ICJ Statute was adopted at the San Francisco Conference that established the United Nations. The Statute “forms an integral part of the [UN] Charter.”Footnote ¹⁰⁵

The Statute authorizes the ICJ to exercise two primary forms of jurisdiction: it hears legal disputes submitted to it by states (contentious cases) and issues advisory opinions (advisory proceedings) on legal questions referred to it by duly authorized UN organs and specialized agencies. The Court does not exercise compulsory jurisdiction over contentious cases. Rather, the Court is competent to entertain a dispute only if the states concerned have accepted its jurisdiction.Footnote ¹⁰⁶ The Court's decisions in contentious cases are final and binding, and the Charter provides that UN Members “undertake[] to comply” with ICJ decisions “in any case to which it is a party.”Footnote ¹⁰⁷

The Court consists of fifteen judges, and normally hears cases en banc. However, a state party to a case which does not have a judge of its own nationality on the bench may choose a person to sit as a judge ad hoc in that specific case,Footnote ¹⁰⁸ meaning that in any particular dispute the court can consist of up to seventeen judges. The bench may not include more than one national of the same state, and “the [Court] as a whole [should] represent[] … the main forms of civilization and of the principal legal systems of the world … .”Footnote ¹⁰⁹ In practice this principle has led to a distribution of membership of the Court among the principal regions of the globe. In recent years, this distribution has consisted of three individuals from Africa, two from Latin America and the Caribbean, three from Asia, five from Western Europe and other states, and two from Eastern Europe, a distribution which corresponds to that of membership of the Security Council. Moreover, as a matter of practice, the Court has always included judges of the nationality of the permanent members of the Security Council.Footnote ¹¹⁰

How has the ICJ addressed the Judicial Trilemma? As with our analysis of the two European courts, we will begin with discussions of the features that the Court's designers and judges have attempted to maximize, and then turn to the third element associated with the Trilemma.

The ICJ, like the CJEU, and unlike the ECtHR, has always been a high accountability court. Candidates for the bench are not directly nominated by states, but rather through a group consisting of the members of the Permanent Court of Arbitration (PCA) designated by that state. Having nominations flow through the PCA, rather than directly from states, was intended to add an independent, professional dimension to the process, and to reduce the political element. This effort, however, has met with only partial success. As one former ICJ judge remarked, “The process is heavily political. Nomination by national groups has not worked; it is meant to be insulated from governments. But governments do make nominations. The national group system is largely ineffective.”Footnote ¹¹¹

Candidate names are submitted to both the UN General Assembly and the Security Council. To obtain a seat on the Court, candidates must receive an absolute majority of the votes in both bodies, which cast their votes simultaneously but separately.Footnote ¹¹² Successful candidates are elected to nine-year terms of office, and are eligible for reappointment.Footnote ¹¹³

The reappointment process is similar to the appointment process; potential candidates must receive the support of their home state, and the candidate is then considered by both the General Assembly and the Security Council. There is no limit on the number of times that incumbent judges can be reelected. Thus, like judges at other international courts, ICJ judges are de facto dependent upon their home state for renomination. Their path to reelection, however, is more difficult than that of CJEU judges (as other EU members typically approve of other state's judicial nominees) but easier than that of WTO AB members (who, as discussed below, must be approved by consensus).

The ICJ has been a high transparency court, although the issue has given rise to controversy from time to time. The ICJ Statute provides that the judgment “shall contain the names of the judges who have taken part in the decision,” echoing language found in the PCIJ Statute. A 1926 revision of the PCIJ's Rules of Court elaborated on this language by providing that the judgment specify the number (but not the names) of the judges in the majority. During debates over rule reforms in 1926 and 1936, the judges considered requiring that the judgment list the names of the judges in the majority, but in both instances a majority of judges rejected the call for greater transparency.Footnote ¹¹⁴

The initial ICJ Rules of Court followed the PCIJ Rules. In practice, the absence of a rule requiring disclosure of how each judge voted, coupled with a rule permitting but not requiring dissents (discussed below), meant that in some cases it was not possible to determine how particular judges voted.Footnote ¹¹⁵ In 1978, however, the judges adopted a significant change to the Rules of Court, providing the judgment contain not only “the names of the judges participating … ,” but also “the number and names of the judges constituting the majority.”Footnote ¹¹⁶ By implication, of course, this rule identifies the number and names of the judges who are not in the majority. Moreover, in 1980, the Court adopted the practice of voting paragraph by paragraph on the dispositif and listing the names of the judges who voted for and against each paragraph, producing even greater transparency.

The PCIJ and ICJ have also witnessed longstanding debates over the desirability of permitting separate opinions, including dissents.Footnote ¹¹⁷ The issue had been highly controversial during the drafting of the PCIJ Statute. In a compromise proposal, an Advisory Committee of Jurists recommended allowing the publication of dissenting votes, but not of dissenting opinions. Thereafter, the League's Council amended the draft PCIJ Statute to permit judges to append a statement of their individual opinions to the Court's judgment. As finally adopted, Article 57 of the PCIJ Statute provided that “[i]f the judgment does not represent in whole or in part the unanimous opinion of the judges, dissenting judges are entitled to deliver a separate opinion.”

The Court's rules went even further, authorizing judges to issue dissenting opinions in advisory proceedings, even though this was not expressly provided for in the court's Statute. Moreover, although the Statute did not provide for concurring opinions, these were introduced as a matter of judicial practice in 1923.

In 1926, during deliberations over revisions to the rules of court, Judges Loder and Weiss proposed to eliminate the rule permitting dissents in advisory proceedings, but the proposal was not accepted.Footnote ¹¹⁸ Another proposal, endorsed by Judges Anzilotti, Finlay, and Moore, pushed in the other direction. It would have provided that the names of all judges who dissented, as well as any opinions they might write, be published. The proposal was addressed, in part, to a judicial practice that had developed whereby “certain judges did not dissent publicly, but filed secret dissenting opinions with the minutes of the Court's deliberations without attaching them to the judgment or the opinion.”Footnote ¹¹⁹ While the court did not make the mention of the dissenting judges' names mandatory, the judges did decide to end the practice of attaching secret dissents to the minutes of the Court.Footnote ¹²⁰

In 1929, another process for revising PCIJ rules was undertaken, and Judge Fromageot proposed to eliminate the ability to dissent in advisory jurisdiction cases. This proposal triggered strong opposition. Judge Hurst replied that the proposal would “destroy the Court;” Judge Root argued that “the suppression of dissentient opinions would … be disastrous;” and Judge Politis noted that, although he originally opposed the idea of dissenting opinions, he had come to see them as being of “immense advantage to international law.”Footnote ¹²¹ Judge Fromageot then withdrew his proposal, and the issue was not formally considered again at the PCIJ.

In the run-up to the San Francisco conference, an informal Inter-Allied Committee proposed that “it should be obligatory on any judge who dissents from the majority to state his reasons for so doing.”Footnote ¹²² However, the proposal was not adopted, and the ICJ Statute echoes the PCIJ Statute in providing that judges may, but are not required to, dissent publicly when voting against the majority decision. The Court's Rules provide that “any judge may … attach his individual opinion to the judgment, whether he dissents from the majority vote or not, or a bare statement of his dissent.”

From the start, PCIJ and ICJ judges have frequently exercised their right to write separately, and it is extremely rare for the Court to release a judgment without separate opinions. For example, in its first 243 decisions (90 judgments, 25 advisory opinions, 128 orders), the Court also released 1,017 individual opinions, including 349 dissenting opinions, 406 separate opinions, and 262 declarations.Footnote ¹²³ The practice of frequent dissent continues. By way of example, the Court's last five judgments in contentious cases have been accompanied by eight,Footnote ¹²⁴ three,Footnote ¹²⁵ ten,Footnote ¹²⁶ four,Footnote ¹²⁷ and twelveFootnote ¹²⁸ separate opinions, respectively.

What implications does high accountability and high transparency have for judicial independence at the ICJ? The Court's Statute provides that “[t]he Court shall be composed of a body of independent judges, … who possess the qualifications required in their respective countries for appointment to the highest judicial offices, or are jurisconsults of recognized competence in international law.” Significantly, the Statute also provides that “[j]udges of the nationality of each of the parties shall retain their right to sit in the case before the Court.” This clause suggests that judges serve in their individual capacity, and it is not unknown for a national judge to vote against his or her own state's legal position.Footnote ¹²⁹

These provisions are reinforced by several structural features designed to enhance independence. For example, ICJ judges “enjoy diplomatic privileges and immunities” when engaged on the business of the Court.Footnote ¹³⁰ And once they assume their positions on the bench, judges are protected from removal. No member of the Court can be dismissed unless, in the unanimous opinion of the other members, she or he no longer fulfills the required conditions.Footnote ¹³¹ In fact, a removal under these circumstances has never occurred.

Are these provisions sufficient to ensure judicial independence—or does the combination of high accountability and high transparency inescapably pose a structural threat to judicial independence? The answer to this question turns, in part, on judicial motivations. On the basis of our interviews with international judges, we believe that, at least some of the time, at least some judges are more interested in discharging their judicial duties and deciding cases as best they can, than they are in securing an additional term on the bench. Moreover, many international judges, particularly those near the end of their careers or with other career options, may care more about their reputation among their peers and the invisible college of international lawyers than they do about reappointment. As Alvarez correctly notes, “[t]here is considerable evidence that those who serve on [international] tribunals, including its judges, see themselves as [Hersch] Lauterpacht would have described them as agents of the ‘international community’ bent on the pursuit of ‘justice’ broadly understood.”Footnote ¹³²

Nonetheless, we assume, as many international judges and scholars of international courts maintain, that nearly every international judge is motivated, at least in part, by a desire to keep his or her position (i.e., to be reappointed at the end of a term), whether as an end in itself or as a means to other ends (i.e., to produce changes in legal doctrine, or to obtain influence or power, or simply to serve). Thus, even if a desire for reappointment is only a secondary or tertiary consideration, the need to stand for reappointment creates pressures, particularly near the end of a judge's term, which can threaten judicial independence. A distinguished scholar who appeared before the Court and served as a judge ad hoc summarized the critique of reelection as follows:

The concerns over the pressures on independence caused by electoral processes are heightened by the fact that successful candidates (and their nominating states) typically undertake lengthy “campaigns,” involving visits to UN delegations by candidates and substantial lobbying by diplomats. For example, in his pursuit of a position on the ICJ, Judge Keith's electoral “campaign … lasted for over two years and involved visits to more than 30 capitals as well as [three trips] to New York. …”Footnote ¹³⁴ The trend toward increasingly lengthy and elaborate campaigning has only intensified in recent years. Mackenzie and Sands note that “[e]lections involving judges standing for reelection can focus on cases decided by the judge,”Footnote ¹³⁵ and diplomatically conclude that “[t]his practice raises many eyebrows.”Footnote ¹³⁶

Court insiders have been considerably less circumspect in describing the structural pressures that can arise from the ICJ's reelection system. One former president judge of the Court complained that “[t]he three-yearly elections often involve a good deal of horsetrading at the best of times; henceforth, this has become accentuated, leading to a situation in which candidates have been and will in future be likely to be elected more on a consideration as to how they might vote on certain delicate issues coming before the Court than on whether they would or could render objectively valid judicial opinions in all cases and at all times.”Footnote ¹³⁷ And drawing on his two terms of service on the Court, Sir Gerald Fitzmaurice observed that:

We read these and other comments from distinguished judges and Court insiders as confirming the underlying logic of the Judicial Trilemma.Footnote ¹³⁹ Of course, recognizing that the Trilemma is keenly felt in the lived experience of those who serve on the Court is not equivalent to stating that ICJ judges will necessarily change their votes or even shade their opinions to curry favor with those who control their renomination and reelection. Nor does describing these pressures impugn the personal integrity of judges, many of whom have demonstrated over the course of their careers that they are prepared to sacrifice personal and professional gain to advance rule of law values. Rather, recognizing and acknowledging the Trilemma reveals that the Court's particular structural features require judges who are committed to judicial independence to resist pressures in ways that are not required of judges at courts with non-renewable terms—and, as we shall see below, that this outcome can be avoided by changing other structural features of the environment in which they operate.

Thus, the ICJ reveals an ambiguous portrait of a third position vis-à-vis the Trilemma. Unlike the CJEU and the ECtHR, the ICJ is a high accountability, high transparency court. Some observers argue that, as the Trilemma would suggest, this combination of structural features produces a Court that is low on judicial independence. This position can be graphically illustrated as follows:

Figure 4.

The Judicial Trilemma at the International Court of Justice

We hasten to underscore that claims regarding the level of judicial independence at the ICJ are highly contested,Footnote ¹⁴⁰ with many respected observers concluding that judicial independence at the ICJ is high. But whatever judgment one ultimately makes about the degree of independence of ICJ judges, there should be little doubt that the structural environment in which they find themselves places pressures on judicial independence that judges at the CJEU and ECtHR do not experience.