The ICJ’s Unwritten Rulebook: Regional Succession and Institutional Equilibrium

For almost eighty years, United Nations Member States have behaved as if a mandate required them to maintain regional continuity in elections to the International Court of Justice. Yet, the Court’s Statute contains no such rule. Article 9 requires only that the Court, taken as a whole, represent “the principal legal systems and civilizations of the world”; it allocates no seats among regions, establishes no geographic quotas, and nowhere requires that a departing judge be succeeded by a candidate from the same regional group. The practice has nonetheless persisted, unchanged, across dramatically different geopolitical conditions.

The historical record reveals an extraordinary pattern of electoral continuity. Of the 154 seat successions that have occurred through regular and special elections since 1946, only eight have resulted in a seat passing from one regional group to another. Seven of those transfers occurred during the period of decolonization, when the regional architecture of the United Nations was fundamentally rearranged. During the past fifty-six years, there has only been one instance of a regional transfer between a departing and a succeeding judge.

Why have Member States consistently reproduced a pattern that international law does not require? This post argues that the repeated practice of regional succession has evolved into an unwritten constitutional convention governing elections to the ICJ. It further argues that the remarkable durability of that convention is best explained by a self-reinforcing institutional equilibrium, sustained through reciprocal expectations among regional groups. Once established, each group acquired a continuing institutional interest in preserving the convention, because respect for another group’s established representation reinforces confidence that its own regional representation will be respected in future elections. Scholarship has long recognized the importance of informal institutions and constitutional conventions in shaping the operation of domestic constitutional systems and more recently, international organizations. Yet comparatively little attention has been devoted to the persistence of unwritten constitutional conventions governing representation in international courts. By offering the first systematic reconstruction of regional succession on the ICJ, this post identifies such a convention and advances an institutional explanation for its remarkable durability.

Institutional background

Article 9 of the ICJ’s Statute sets out the only textual guidance Member States have on the composition of the bench. The language regarding representation is noticeably abstract. It names no regions, allocates no seats, and imposes no regional quota to the composition of the bench. It stands in sharp contrast to the UN Charter’s language used to define the Security Council, whose five permanent seats are fixed by name in Article 23, and whose non-permanent seats have, since 1963, been subject to an explicit geographic distribution formula adopted via General Assembly resolution 1991 (XVIII).

The ICJ’s electoral mechanics are layered and demanding. Fifteen judges are elected for nine-year terms, with five seats coming open roughly every three years. Candidates are nominated not by governments directly but by national groups. To be elected, a candidate must secure an absolute majority in both the General Assembly and the Security Council, who vote simultaneously but separately. Because the two bodies must independently converge on the same candidate, a nomination that prevails in one chamber but does not in the other can produce drawn out multi-round elections. Nothing in this formal process requires that a departing judge be replaced by a candidate from the same regional group. In practice, however, member states have organized themselves into five distinct regional groups: the African Group, the Asia-Pacific Group (APG), Latin American and Caribbean Group (GRULAC), Western European and Others Group (WEOG), and the Eastern European Group (EEG). The absence of that statutory basis for this regional organization is what makes the empirical pattern presented in the next section so puzzling: Member States have produced, through informal practice, a degree of regional stability in the Court that is not explicitly mandated in its founding Statute.

The empirical picture

The reconstruction reveals a strikingly simple and remarkably stable pattern. Although the ICJ elects judges in triennial batches of about five, rather than seat by seat, this study reconstructs individual seat lineages by matching each incoming judge to the outgoing judge whose term they succeeded. Therefore, a seat is coded as regionally continuous when the incoming judge’s regional group matches that of their predecessor. Nine of the Court’s fifteen seats have never changed regional affiliation since 1946, as seen in Figure 1, a visual rendering of these seat lineages adapted from a larger dataset covering every ICJ election. These include the seat continuously occupied by successive French judges, the seat continuously occupied by successive American judges, two additional WEOG seats, two GRULAC seats, one Asia-Pacific Group seat, and two Eastern European Group seats (including the seat formerly occupied by the Russian Federation, which despite Russia’s electoral defeat in 2023, remained within the EEG through the election of Romania).

Figure 1. These nine ICJ seats have been held by the same regional group since 1946.

The remaining six seats form what might be called the heterogeneous group: seats whose regional affiliation has changed at least once since 1946. It is within this smaller group of six seats that all eight recorded switches occurred.

Figure 2. These six ICJ seats have changed regional groups since 1946.

Two features of this table are worth noting before turning to the explanation in the next section. First, the switches are not evenly distributed across the eighty-year period; they cluster overwhelmingly in a twenty-year window. Second, every election after 1970, with the single exception of 2018, continues with the norm of regional continuity for close to half a century. This concentration is graphically shown in Figure 2: dense color transitions in the 1950s and 1960s give way to nearly fifty years of uninterrupted single-color bars.

Explaining the regional switches

The concentration of switches between 1952 and 1970 is not a coincidence of timing; it is the direct institutional consequence of decolonization’s effect on United Nations membership. The scale of that effect is difficult to overstate. As David Kay observed in his 1967 study of the period, The Politics of Decolonization: The New Nations and the United Nations Political Process, only 13.2 percent of UN Member States in 1955 were states that had achieved independence since 1945. By the end of 1966, that figure had risen to 45 percent (Kay 1967). This influx did not merely add votes to the General Assembly; for the first time it created a coherent Afro-Asian voting bloc. The ICJ’s regional switches of this period should be read against this same institutional backdrop. The Africa Group did not exist as a recognized electoral bloc in 1946, as at that time there was no critical mass of independent African states to constitute one. By the mid-1960s that had changed entirely, and the six switches in Figure 2, dated between 1952 and 1970, largely represent the practical accommodation of this new bloc within the Court’s existing seats. The Africa Group’s seats on the Court were carved substantially from allocations that had previously been considered WEOG or GRULAC seats, mirroring the same renegotiation of institutional geography that produced the Security Council’s own enlargement in the same years. Once this reallocation was substantially complete by 1970, the pattern froze. For the following forty-eight years, every seat passed within its regional group without contest.

The 2018 election broke this near half-century old pattern. In the final rounds of voting, Dalveer Bhandari of India prevailed over the United Kingdom’s Christopher Greenwood, both incumbents, in a contest that ran to five rounds. This was the first instance in the Court’s history in which a sitting judge from a permanent Security Council member lost re-election, and the first instance in which the United Kingdom lost its seat on the ICJ. What makes this election analytically significant is not merely that a P5 state lost a seat, but how the seat was reallocated once it was lost: the vacancy passed cleanly from WEOG to the Asia-Pacific Group through the Court’s ordinary electoral mechanisms, rather than through any broader unsettling of the regional structure. The Court’s subsequent 2023 election returned to the pattern of uncontested regional succession: every departing judge was succeeded by a candidate from the same regional group. During this election, the Russian Federation, a P5 state, lost a seat in the ICJ only to be replaced by a Romanian of the same regional group.

What the pattern tells us

The central empirical finding of this post is that regional continuity on the ICJ is neither a historical coincidence nor merely the product of political habit. The evidence strongly suggests that over successive generations of elections, Member States have consistently reproduced the expectation that seats ordinarily remain within the same regional group, despite the absence of any legal rule requiring it. Once the expectation became an established practice, each regional group acquired a continuing institutional interest in preserving it. Respect for another group’s established representation on the bench reinforced confidence that one’s own regional representation would also be respected in future elections. As a result, compliance depended not upon legal obligation, but upon the shared expectation that maintaining regional continuity served the long-term interests of every regional group. However, the distinction between the observed practice and its proposed explanation is important. The constitutional convention is the empirical phenomenon documented by this study, and institutional equilibrium is the mechanism that best explains why that convention has endured for almost eighty years.

This explanation also helps account for the historical evolution of the practice, having been shaped by the period of decolonization. The emergence of newly independent African and Asian States fundamentally altered the political composition of the United Nations and required a renegotiation of the Court’s regional balance. Seven of the eight regional reallocations identified in this study occurred during this period of institutional adjustment. Once that redistribution had largely been completed by the early 1970s, the pattern stabilized and regional succession resumed as the ordinary expectation of ICJ elections. The same analytical framework helps explain the significance of the 2018 election. The transfer of a seat from the Western European and Others Group to the Asia-Pacific Group represented the first regional reallocation in almost half a century. Despite this, the broader convention remained intact.

But the case also raises questions that extend well beyond what this paper can resolve. If an unwritten constitutional convention has governed the composition of the world’s principal international court for almost eighty years without ever being codified, then the role of constitutional conventions in international institutional law may be considerably greater than existing scholarship has generally recognized.