On 1 January 1995, the World Trade Organization (WTO) replaced the General Agreement on Tariffs and Trade (GATT) as the chief organization governing global multilateral trade. Seventy-six of 128 existing state parties to the GATT directly became WTO members; another thirty-six countries joined the WTO during the following year.Footnote 1

The socalled single undertaking through which the WTO replaced the GATT has been described as a remarkable example of “coercive bargaining”, “raw use of power” or even a “coup against pluralism” by dominant Western states (Steinberg, 2002:342; Jupille et al., 2013; Howse & Langille, 2023:31). However, the GATT’s replacement is far from unique. The Organization for European Economic Cooperation (OEEC) was replaced by the Organization for Economic Cooperation & Development (OECD) in 1961; the United International Bureau for the Protection of Intellectual Property was succeeded by the World Intellectual Property Organization (WIPO) in 1970; and the Conference on Security & Cooperation in Europe (CSCE) gave way to the Organization for Security & Cooperation in Europe (OSCE) in 1990—to name just a few similar cases. Overall, some sixty IGOs created since 1900 have organizational predecessors from which they have directly taken over their mandates, members, and core functions (Eilstrup-Sangiovanni, 2020). Similar to the GATT/WTO succession, in each of these cases an existing IGO was terminated with the specific aim of introducing a new organization that fulfilled a similar function as the discarded incumbent.

Given the cost and uncertainty involved in creating international institutions (Keohane, 1984), institutional replacements present a puzzle: Why would states dissolve an existing IGO only to instate a new one that takes over all (or most) of the former organization’s legal assets, mandate, and functions—a practice labelled by institutional lawyers as institutional succession? (Schermers & Blokker, 2003). Specifically, why would states replace an existing IGO rather than reform it? After all, the 1947 GATT was revised through eight rounds of multilateral trade negotiations before eventually being replaced by the WTO in 1995.

Understanding the conditions in which state parties replace an international institution is important for our understanding of the broader life cycle of IOs (Gray 2018, 2024) as well as giving nuance to dynamics of institutional death (Eilstrup-Sangiovanni 2022, 2021). We explain institutional succession (hereafter “Succession”) as the result of a bargaining process through which dissatisfied parties seek to alter the institutional status quo. Succession is one among several strategies of institutional change: others include reform, shifting an issue from one existing institution to another (known as “regime-shifting”, see Helfer, 2004), or creating new rival institutions to challenge the status quo (“rival regime-creation”, see Morse & Keohane, 2014; Urpelainen & van de Graff, 2014; Vabulas & Snidal, 2013). Whereas reform, regime-shifting and rival creation have attracted much scholarly attention, succession has been generally overlooked, subsumed under the wider category of institutional creation.Footnote 2

To the extent it addresses institutional succession (which is only indirectly as generic “creation”), existing literature portrays succession as inherently more difficult and costly than reform, given that it involves crafting de novo institutions with all the added costs and uncertainties this entails (Cottrell, 2009; Panke & Petersohn, 2011; Jupille et al., 2013). In contrast, we argue that reform and succession are equally expedient mechanisms of institutional change, but under different conditions. On the one hand, by moving negotiations outside an existing institution, succession bypasses veto players that may stunt reform. On the other hand, succession suffers from potential scale diseconomies since not every member of an existing IGO may join its successor, resulting in a loss of efficiency for institutions subject to scale economies (which are the focus of our analysis). By contrast, reform processes typically involve all existing parties to an agreement and are therefore less imperilled by diseconomies of scale. Depending on which negotiation hurdle prevails, we expect that reform will be preferred to succession or vice-versa. For institutions not subject to scale-economies, rival regime-creation presents an alternative mechanism of change (as we illustrate later). Rival creation is not, however, a direct focus of our theory which concentrates on explaining the choice between reform/succession in institutions subject to scale economies.

Our discussion advances choice-based theories of institutional bargaining in three directions. First, we advance existing work on institutional contestation and change by identifying a hitherto neglected mechanism of institutional change—succession—and specifying under what conditions it can be expected. Second, we distinguish succession from other outcomes of institutional contestation such as rival regime-creation. Third, we nuance existing notions of “exit-power” (Hirshman, 1972, Gray and Slapin 2013) and “go-it-alone” power (Gruber, 2000) by showing how what we label “outside options” (a form of collective exit power) can be successfully invoked by coalitions of stronger and weaker states alike.

The article is organized as follows. After a brief review of existing literature in Section 1, Section 2 defines institutional succession and outlines our theorical argument regarding when succession will be preferred to reform, or vice-versa. Section 3 offers four brief cases illustrating our main analytical findings. Three cases (the succession of the International Sugar Council by the International Sugar Organization, 1967; the succession of the GATT by the WTO, 1995; and the succession of the International Vine & Wine Office by the International Organization of Vine & Wine, 2001) are explicitly selected because they display the dependent outcome of interest: succession. However, each organization has also undergone long periods of continuity and several instances of reform prior to being replaced, thereby ensuring within-case variation on the dependent variable. In addition, we consider one “control” case, the IMF, which has been reformed several times but never replaced. By demonstrating the consistency of within-case and across-case findings we strengthen confidence in our theory (Mahoney, 2003:360–2). We do not claim this research design offers a full test of our theory. Rather by illustrating how different mechanisms of institutional change unfold according to our expectations across a diverse set of cases we confirm the plausibility of our theory and to identify potential avenues for theoretical refinements to guide future research. Our case analysis is based on both primary sources (treaties, protocols, diplomatic cables, minutes of meetings, and private papers of senior international bureaucrats) and secondary academic sources.

1 Existing literature

Why do states sometimes dispose of an existing IGO only to replace it with a new organization that fulfils an equivalent function? Existing literature offers limited insights on this question. For many decades, institutionalist scholars of all theoretical stripes cited high costs of negotiation and contracting (Keohane, 1984; Stein, 1990); self-reinforcing lock-in mechanisms (Stinchcombe, 1965; Pierson, 2004); learning effects (Pierson, 2004; North, 1990); socialization (Finnemore, 1993; Cottrell, 2009), and cognitive biases such as endowment effects and bounded rationality to ground the expectation that states prefer to “stick with the institutional devil they know” rather than embark on new and uncertain cooperation ventures (Jupille et al., 2013:7; 2017:118).

More recently, scholars have theorized institutional choice under conditions of “institutional complexity”—when multiple separate, but partially overlapping, institutions exist in a governance domain (Eilstrup-Sangiovanni & Westerwinter, 2021). Some have focused on how powerful states may (threaten to) shift a contested issue from one existing institution to another where they are more likely to get their way (Helfer, 2004; Urpelainen & van de Graaf, 2014). Others have focused on how dissatisfied states may seek to challenge the status quo by creating new, rival, organizations that will become focal points for cooperation (Morse & Keohane, 2014:387; Urpelainen & van de Graff, 2014; Vabulas & Snidal, 2013; Kellerman, 2019). Unlike such strategies which lead to the existence of parallel, overlapping institutions with potentially inconsistent rules and procedures,Footnote 3 succession directly replaces an existing IGO with a formal successor that takes over the former IGO’s membership, mandate, and core functions, leaving just one institution in place.

Since it involves the creation of a new institution to replace an existing one, succession is a priori presumed to be costly and difficult (Cottrell, 2009; Jupille et al., 2017:127). Thus, the only study we are aware of that theorizes institutional replacement relative to a wider set of strategies for institutional change, by Jupille et al. (2013, 2017), invokes bounded rationality to argue that continued use of existing institution is always the default choice, whereas states will consider creating a new institution “only when all other options [i.e. living with the status quo, regime-shifting, or reform] are exhausted” (2013:7, 39; 2017:126–7). By contrast, we argue that, in the hands of fully rational actors, reform and succession present equally suitable instruments of institutional change, but under different circumstances, which we theorize in the next section.

2 A theory of institutional change

2.1 Key concepts

Institutional change can take many forms; from reform to regime-shifting, rival regime-creation, or complete dissolution whereby states return to pre-institutional non-cooperation. Here we focus on two main forms of institutional change: reform and succession, although our argument also has implications for other forms of change such as rival creation.

Reform refers to the amendment of an institution’s founding treaty to introduce new rules or procedures or amend existing ones. Reform is provided for by most IGO founding treaties and typically requires either unanimity or some form of super-majority. As such, reform processes generally include all member states.Footnote 4

Institutional succession occurs when an international institution along with its implementing body is formally dissolved and replaced by a successor that assumes all (or most) of its functions (Myers, 1993; Schermers & Blokker, 2003:145). Importantly, succession implies an existing IGO terminating as a direct consequence of a new one coming into force and formally replacing it, as was the case when the Organization for African Unity gave way to the African Union on 9 July 2002. Succession thus entails institutional dissolution and creation through a single undertaking, with one conditioned upon the other.

As a legal act, succession applies only to formal IGOs constituted by charter or treaty, not to informal norms, conventions or customs which can only be informally replaced. Succession must similarly be distinguished from situations in which a formal IGO is informally replaced by a competitor that fulfils a similar purpose (Schermers & Blokker, 2003).Footnote 5 For example, the creation of the Organization of African States (OAS) in 1963 resulted in the gradual disuse of the Conference of Independent African States (Wessels, 2011). Yet, since there was no direct transfer of functions from one to the other, this is not a case of succession, but rather of organizational rivalry leading to the gradual demise of the Conference and the eventual assimilation of its functions by the OAS.

Finally, succession differs from rival regime-creation (Morse & Keohane, 2014) since—unlike rival creation—succession does not give rise to parallel, competing institutions, but cleanly wraps up an existing IGO before unfolding a new one. This is a crucial distinction, and one that dictates an important scope condition for our theory of institutional choice. If there were no significant loss of efficiency from operating parallel, rival institutions, states that found themselves dissatisfied with an existing institution could presumably create a more auspicious competitor at relatively low cost. However, many international institutions are subject to significant scale-economies and network effects: the more members, the more efficient the institution is (Meyer, 2010). Where scale-economies obtain, the existence of multiple competing institutions reduces aggregate payoffs from cooperation while increasing overall transaction costs. Commodity agreements are perfect examples, for—unlike, say, development aid which can be pursued effectively through separate small-scale initiatives (Lipscy, 2017:28)—there is room for only one effective cartel at a time.Footnote 6 Scale-economies are also present in most regulatory regimes, such as the various “Basel” committees. In areas such as human rights, refugee protection, or peacekeeping, the need for legitimacy often militates against competition between rival organizations, although it does not always exclude it. We posit scale economies as a scope condition of our theory of succession (condition #1).

When successful, succession may deliver sweeping institutional change—such as abolishing some institutional functions or reshaping membership—that exceeds what can typically be achieved through reform (Klabbers, 2002:325). The replacement of the GATT provides a good example. However, sweeping change is not a defining feature of succession. As we show in the next section, the fundamental difference between reform and succession is in the instrument of institutional change; either a (super)majoritarian decision by existing state parties in the case of reform, or, in the case of succession, an extra-institutional bargain initiated by a subset of members who, by moving negotiations outside an existing institution, can present a new agreement as a “take-it-or-leave-it” deal to others. Reform and succession represent different mechanisms of institutional change—not different degrees of departure from the status quo.Footnote 7

2.2 Dynamics of institutional change

Since they arise from consensual decisions by sovereign states, we assume that, at the point of creation, IGOs deliver actual or expected benefits for all contracting parties. However, subsequent events (endogenous or exogenous to institutions) may produce winners and losers, leading to demand for institutional change.

Our theory of institutional change builds on a stylized version of an institution as comprising two coalitions: institutional leaders is the coalition that sets the agenda, while institutional followers can either reject or accept the measures proposed by institutional leaders. The notion that there are only two coalitions is a standard simplifying assumption when modelling institutional bargaining processes.Footnote 8 Similarly, agenda control is a simple, yet realistic, way of conceptualizing institutional power. Agenda-setting power (the formulation of proposals that are difficult to amend) is sometimes a formal right of powerful states, sometimes a de facto capacity of resource rich actors to control the parameters of discussion and limit proposals for consideration (Bachrach & Baratz, 1962; Steinberg, 2002:341; Lukes, 2005). Weaker states, by contrast, are generally agenda-takers.Footnote 9 This does not mean they are powerless. While they may not hold individual veto power, insofar as reform typically requires super-majoritarian support, institutional followers can often block agreement until their views are considered (Eilstrup-Sangiovanni, 2023; Hofmann, 2019).

In an institutionalized context of this kind, initial payoffs from cooperation tend to reflect this inherent power discrepancy in that institutional leaders offer terms of cooperation that yield low payoffs to followers—roughly equal to the latter's reservation value (that is, what they would receive absent an agreement). Leaders then claim the residual benefits from cooperation. The deal is unequal, yet stable because, their reservation being met, there is nothing followers can do to improve upon it.

The institution is stable until an unexpected change in circumstances undermines the initial deal. We assume that demand for institutional change always reflects an unexpected change in circumstances. If the change was expected, we would expect institutional founders to have anticipated how these circumstances would generate demand for institutional change and to have designed the institution to insure against such circumstances from the outset. To introduce a choice between reform and succession, the adverse effects of changing circumstances must also be asymmetrical. We could imagine an exogenous shock that reduces cooperative payoffs to all parties lest some technical fix is implemented. In this case, reform would presumably be straightforward to agree per the standard assumption that cooperation is relatively easier when distributional differences are outweighed by joint efficiency gains (Fearon, 1998; Krasner, 1991). An unexpected and asymmetrical shock is thus a second theoretical precondition for succession to occur (condition #2).

2.2.1 Inside v. outside options

Assuming the initial payoff distribution from cooperation is disturbed by an unexpected and asymmetrical change in circumstances, what can the losing side do to rebalance payoffs? If institutional leaders are the losers, they will typically prefer to uphold the initial deal because it provides them with sufficient benefits above their reservation value to absorb a loss and still gain from cooperation (and because maintaining the status quo is likely to be preferable to whatever reforms would be accepted by followers).Footnote 10 As we explain below, the main reason why founders would pursue institutional change is therefore in response to followers suspending or decreasing cooperation. As a rule, institutional leaders are typically institutional defenders.

The case of institutional followers is different: Earning the expected equivalent of their reservation value, even a small reduction in payoffs will make them indisposed to continue cooperating under an existing agreement and turn them into institutional challengers. They have two alternative strategies available: an inside option and an outside option. An inside option is a strategy (and associated payoffs) that a party can use to improve their final payoffs while the parties temporarily disagree (Muthoo, 1999:137). The textbook illustration is a union going on strike during wage negotiations with a firm; each new day spent striking reduces both sides’ value for the game and increases pressure to come to an agreement. In contrast, an outside option is the payoff to a player if she chooses to permanently stop bargaining and chooses not to reach an agreement. Inside options in the context of IGOs involve suspending or decreasing cooperation in ways that reduce collective gains, whereas outside options involve abandoning negotiations, possibly to pursue an alternative agreement instead (Lipscy, 2017).

Inside options in the context of bargaining over institutional change can take three forms: ‘Foot-dragging’ reduces joint benefits from cooperation by only partially implementing agreed-upon deals, or through stalling agreement on new initiatives (Eilstrup-Sangiovanni, 2023). A second inside option is to invoke an escape clause or a safeguard measure (say, antidumping or countervailing duties in the WTO). When implemented, such measures have the effect of weakening a member state’s commitment (Downs & Rocke, 1995; Rosendorf & Milner, 2001; Johns, 2014). Regime-shifting presents a third inside option on our theory because shifting cooperation on a contested issue from one forum to another can serve, in Helfer’s words (2004:42), to introduce "strategic inconsistencies" which imply de facto reneging on an initial agreement.

Irrespective of what form an inside option takes, by resorting to it, Challengers reduce their losses from adverse circumstances. At the same time, they also inflict losses on Defenders. Hence Challengers often invoke inside options to pressure Defenders to renegotiate an agreement. As noted, this pressure is the main reason why institutional Defenders, as the chief beneficiaries of an existing institution, may consent to institutional change—whether through reform or succession.

2.2.2 Stasis or change?

If institutional Challengers exercise their inside option, Defenders need not concede formal institutional change, but may simply accept the new status quo. In this scenario, cooperation continues with no formal change in institutional rules or procedures, but with persistent reneging by Challengers. This outcome (Stasis) may be preferable to Defenders if reneging by Challengers inflicts only limited losses on them. Hence, a third theoretical condition for succession (and indeed for reform) is that Challengers must be able to inflict significant harm on Defenders by invoking their inside option, amplifying pressure for change (condition #3).

A second response by Defenders is to offer a reform that redistributes benefits in favour of Challengers. If successful, reform puts an end to Challenger’s reneging and reboots cooperation on different terms. However, reform may be unachievable, or unappealing, usually due to a split among Defenders, or because one or more disgruntled Challengers condition their consent on excessive demands. The only way to proceed is then to sink the current organization and charter a successor, or, as Steinberg (2002:349) puts it in the context of the GATT/WTO, to collectively ‘withdraw from the deadlocked organization, step into anarchy, and reconstitute a new organization under different terms.’ If successful, this response, whereby either Challengers or Defenders invoke an outside option, delivers Succession. Importantly, on our model, this outside option is not chosen as a last-ditch response to a failed reform process (cf. Jupille et al., 2013) but rather due to its higher expected payoff. For well-informed, rational actors, both mechanisms of change (reform and succession) are in principle always on the table. Next, we explain when either option is likely to be exercised, and by whom.

2.3 Avenues for institutional change: Reform v. succession

Assuming sufficient pressure for formal institutional change, when should we expect reform v. succession? As noted previously, the fundamental distinction between reform and succession is the process through which institutional change is achieved. Reform typically requires some form of supermajority, or even unanimity. In contrast, succession can be initiated by a sub-group who present a new deal as a take-it-or-leave-it offer to others.Footnote 11 If enough others take the new deal, a new institution replaces the old.

Since it demands widespread consent, reform requires the two sides in an institutional bargain to come to a new agreement on how to divide gains from cooperation. Following an adverse change in circumstances, such agreement will typically be reached by Defenders offering Challengers the equivalent of the latter’s new reservation value (i.e., the least favourable division of benefits they are willing to accept given changed circumstances) and pocketing the residual. Getting agreement requires each coalition to act as a single player, much like the case of collective bargaining between a firm and a union.Footnote 12 Often, however, Defenders will be divided between a core group (which we define as the smallest group capable of organizing outside the existing IO), and a larger group who share the broad interests of the core group, but also have specific interests which may lead them to hold up agreement in search of rents (Tsebelis, 2003). The Challenger coalition may likewise be plagued by internal divisions if one or more Challengers condition their agreement on the satisfaction of an extreme demand. The relative cost of reform thus depends on the reservation value that must be met by Defenders, whereas the ability to strike a deal depends on the unity of the two negotiation coalitions—i.e., on the ability of either side to stare down veto-players within their ranks.Footnote 13

The alternative to reform is succession which implies that a core group of either Defenders or Challengers propose a new institution to be individually accepted by other states. Succession has one major advantage over reform; it is not hostage to veto players since it does not require collective acceptance by any coalition. Instead, a core group proposes a new organization and other states individually decide whether to join. Joining the new organization and leaving the old is done in a single undertaking. If enough states decide to join the new organization and sink the old, succession is accomplished.

However, succession has limitations of its own: it only works if enough members are willing to leave the incumbent IGO and join the successor. Again, this is because we posit that institutions liable to succession are subject to scale economies; the more members, the more efficient the institution is (condition #1). An implication of scale economies is that only one IGO of a certain kind is viable: the successor organization either draws in a majority of existing state parties or does not materialize. This is the institutional equivalent of a “tipping game” in game-theoretical terms (Schelling, 1978).

The need to reach a minimum participation threshold (or “tipping point”) is a liability for architects of succession. Assuming that a successor IGO is proposed by a core group, this group is either large enough to tip the scales in favour of the successor on their own, or—if not—must sway enough other members’ allegiance away from the incumbent IGO by offering a more attractive deal (or individual side-payments). Since these prospective members’ decisions are made individually rather than as a coalition, they do not fully internalize the positive externalities that derive from scale-economies. One reason is uncertainty over who will join. Given scale-economies, it is rational for every willing co-operator to join a successor IGO once a critical group of other parties do so. Recall, however, that unlike negotiations over reform, negotiations over succession unfold in a non-institutionalized bargaining setting where information is less abundant. Unlike amending an existing treaty, accepting a new founding treaty may require domestic ratification, introducing an element of uncertainty. Thus, individual states may hold out (or demand robust side-payments) if they doubt a critical group will materialize, making it costly for the core group of proposers to flip enough participants. The upshot is that the smaller the core group that proposes succession is, the less likely that group will succeed in persuading enough others to abandon the incumbent IGO and adopt the new. Succession demands a powerful core group capable of organizing outside the incumbent institution and with sufficient resources to sway other members individually. However, assuming enough other states join the successor and the tipping point is reached, the situation changes; rather than facing a choice between the status quo and succession, “holdouts” now face a choice between joining the successor (on terms dictated by the core group) or exiting the regime all together.

The assumption of uncertainty about whether a tipping point will be reached warrants brief discussion. We assume fully rational actors operating in an information rich environment. This is a key difference to leading theories of institutional choice which invoke bounded rationality and information scarcity to ground an assumed penchant for the institutional status quo (e.g., Jupille et al., 2017:199; Fioretos, 2017). Assuming plentiful information makes sense given the backdrop of an existing IGO facilitating cooperation, which means individual preferences and collective win-sets are relatively well known. Nevertheless, there is always an element of uncertainty arising from domestic unpredictabilities. Effectively, the further the proposing core group is from reaching critical mass, the greater the element of uncertainty as to whether a tipping point can be reached.

2.4 Who initiates succession?

The bargaining dynamic leading to institutional change can be summarized as follows: In the wake of an unexpected shock that adversely affects one side in an institutional bargain (condition #2), painful reneging by institutional Challengers increases pressure for institutional change (condition #3). If Challengers and Defenders are both relatively united (capable of staring down veto-players within their ranks), the sides are likely to be able to agree on reform (assuming meeting the reservation point of the other side is preferable to ceasing cooperation altogether). In contrast, if either coalition is split, this will drive up the cost of reform. In this case, a core group from either side may instead propose to replace the contested institution with a successor to which willing co-operators can accede afresh.

Who initiates succession—Challengers or Defenders—is mainly a function of which side has the stronger core group, meaning a group that (i) is sufficiently unified to reach agreement outside the existing institution, and (ii) commands sufficient resources to sway other members to their side. Depending on who initiates a succession, we refer to succession as either “defensive” if in the hands of Defenders, or “offensive” if initiated by Challengers. The main difference is that Defenders of an existing institution will typically initiate succession to bypass outlying veto players that prevent necessary reforms to end reneging by Challengers, whereas Challengers typically initiate succession to replace an institution that is inherently biased against them.

2.5 Summary and Hypotheses

Theorizing institutional choice at a relatively abstract level, we explain the choice between ‘stasis’, ‘reform’ and ‘succession’ as a function of exogenous shocks and the relative capacity of defending and challenging coalitions to set the agenda and shape choices. Our theory rests on three theoretical preconditions: First, to be subject to succession, an institution must exhibit scale-economies which marshal against the creation of parallel, functionally equivalent IGOs (condition #1). Second to undergo formal change, an institution must be subject to an unexpected and asymmetrical shock which leaves a subset of parties dissatisfied with the status quo (condition #2). Third, to build pressure for institutional change, Challengers must be able to inflict losses on Defenders by reneging on the original agreement (condition #3). When these conditions are satisfied, our theory yields the following hypotheses regarding the direction of institutional change:

  • H1: If the coalitions of institutional Defenders and Challengers are both unified (capable of closing ranks and staring down veto-players) reform is likely.

  • H2: If either side is divided by veto players that stunt reform (either by increasing Challengers’ reservation point or by reducing Defenders’ willingness to make the necessary concessions to end reneging by Challengers), succession is likely. In this case who initiates succession (Defenders or Challengers) depends on who has the most viable (i.e., the larger and more powerful) core group.

  • H3: If BOTH coalitions are led by weak core groups, the likely result is either no change (‘stasis’) or, if the cost of perpetuating the status quo are deemed too high, organizational dissolution.

The rationale for H1 is straightforward. Reform requires cohesion on both sides due to the unanimity or quasi-unanimity rules that govern reform procedures in most IGO charters.

The rationale for the two succession outcomes is also straightforward: it takes a strong and resourceful core group to pursue succession. The less united and resourceful the other side is, the easier for the core group to impose their preferred version of succession (which, unlike reform, detractors cannot block through a formal veto). Who initiatives a succession has likely implications for subsequent institutional design. Given that Succession arises from an unwillingness by either a challenging or defending coalition to accept the status quo, we expect the design of a successor IGO to clearly favor the proposing coalition, for example, by introducing a mandate that is more closely aligned with their interests, and/or by structuring voting rules in a favorable manner.

Last, absent group unity or a viable core group on either side, formal institutional change fails, meaning that either the status quo prevails, or the organization is dissolved. The former outcome (stasis) implies that either Challengers accept negative payoffs or Defenders tolerate widespread reneging by Challengers (Table 1).

Table 1 Institutional change outcomes

In the next section we provide empirical support for these predictions. Before we introduce our empirical cases, a caveat is warranted. Our model assumes the centrality of states in driving institutional change. This assumption has been questioned by scholars emphasizing the role of IGO secretariats in ‘masterminding’ institutional change (e.g., the GATT Secretariat is said to have played an important role in proposing blueprints for the WTO). But whilst secretariats may propose reforms or produce blueprints for succession, both forms of change must ultimately be approved by member states. Thus, regardless of who proposes change, the outcome will depend on the relative preferences and resources of institutional Defenders and Challengers.

Once again, unlike institutional choice theories in which boundedly rational actors prone to satisficing contemplate costly institutional choices such as reform or creation only once prior strategies (i.e., continued use or regime-shifting) fail to provide “good enough” solutions (Jupille et al., 2013:199), our theory assumes that fully rational actors consider all available strategies simultaneously. This does preclude that different strategies of change may unfold sequentially, but this is due to different strategies presenting optimal solutions at different times, not due to an inherent inclination to make limited and gradual moves away from the status quo.

3 Empirical illustrations

In the absence of a comprehensive dataset on institutional changes to IGOs, we illustrate the key outcomes predicted by our theory with four brief cases. Cases are chosen according to criteria set by our theoretical scope conditions which narrow the universe of cases to which our theory applies: First, an institution must be subject to scale-economies which marshal against the creation of parallel, functional equivalent IGOs (condition #1). Second an institution must be subject to an unexpected and asymmetrical shock that leaves a subset of parties dissatisfied with the status quo (condition #2). (Recall that an unexpected shock is a precondition for formal institutional change on our theory, whereas an asymmetrical shock is necessary to introduce an element of distributive conflict without which we assume reform to be relatively straightforward).

Beyond these two criteria we chose cases with a view to ensuring maximum variation on our dependent variable (both across and within cases) and thereby illustrate all possible outcomes covered by our theory. We first introduce three cases of succession. Two of these—the succession of Vine & Wine Office (2001) and of the GATT (1995)—are defensive successions triggered by institutional Defenders’ desire to end hurtful reneging by Challengers. The third case—the succession of the International Sugar Council—is a case of offensive succession, initiated by institutional Challengers. Within each case of succession, we also consider previous instances of reform to leverage within-case variation. We further illustrate the reform outcome with the case of the IMF which has undergone several reforms, but, we argue, is unlikely to experience succession. Our focus is on the key moments of institutional choice that illustrate our theoretical concepts rather than on providing a detailed history of each IGO. While a wider range of empirical cases could further enhance confidence in our theory, the fact that it accounts for institutional outcomes in a diverse range of empirical cases attests to its strong plausibility (Table 2).

Table 2 Overview Cases

3.1 The succession of the international office of vine & wine

Faced with overproduction and falling prices, in 1924 France and seven other European wine-producing countries created the Office International du Vin (OIV) to set common standards for production and trade of wine, based on the French Appellation d'Origine Contrôlée (AOC) system which designates wines according to the geographic area (‘terroir’) and manufacturing methods by which it is produced (Hannin et al., 2006:75; Simpson, 2009:68).

During the late twentieth century, as new wine producing countries emerged both within and outside Europe, the OIV adjusted to new market players and new production technologies which led to a doubling of global wine production, and a six-fold expansion of international trade in wine.Footnote 14 During this period, the OIV increased its membership (from 8 in 1924 to 47 in 1990) and amended many of its regulations. It also adapted to changing conditions through flexible application of existing rules. For example, the US joined the OIV in 1984 with a reservation on the application of AOC standards. Despite these adjustments, in 2001 a majority of members decided to dissolve the OIV and replace it with the new International Organization of Vine & Wine (IOV).

A major source of dissatisfaction with the incumbent OIV was an unexpected shift in the balance of market power (condition #2) to which the organization failed to adapt (Hannin et al., 2006:79). During the 1980s, European countries which had traditionally accounted for close to three-quarters of global wine production began to rapidly lose market shares to “new world” wine countries (NWCs). Thanks to more modern technologies as well as scale-economiesFootnote 15 NWCs could produce consistently high-quality wines at lower cost than “old wine countries” (OWCs) (Meloni & Swinnen, 2013). Yet this changing balance of market power failed to be reflected in the OIV’s governance structure which allocated a majority of votes and power of initiative to OWS. During the 1980s, NWCs (the institutional Challengers of our story) grew increasingly dissatisfied with the OIV’s insistence on terroir as the basis for wine classification; they demanded reform (New Zeeland MFA, 2001; Hannin et al., 2006:78). As a result, members divided into two coalitions: on one side stood OWCs who defended the institutional status quo; on the other side stood NWCs (comprising South Africa, Australia, New Zealand, Chile, Argentina, Canada, and the US) who demanded institutional change (New Zeeland MFA, 2001).

Minor reforms were agreed in the late-1980s between the two coalitions. However, these failed to satisfy Challengers, as OWCs, led by France, refused to abolish core restrictions on wine production and classifications. Feeling the negative impact of these market restrictions more and more as their production capacity continued to grow, by the mid-1990s, NWCs invoked their inside option by threatening to shift discussions on deregulating wine trade to the GATT. This regime-shifting move was facilitated by the Uruguay Round Final Act of 1994 which threatened to undermine the legitimacy of several existing OIV regulations. Specifically, the OIV’s ban on the use of oak chips (popular among NWC winemakers) and strict controls of the geographic origins of grapes were seen to be at odds with new GATT/WTO rules which sought to limit technical trade barriers (Hannin et al., 2006:85).

In response to heightened pressure for reform (generated by the prospect of regime-shifting), a review conference was called in 1997 to discuss how to make OIV standards more compliant with WTO practices. NWCs, backed by major wine importing countries such as the UK, pushed for strong deregulation, whereas European wine producers offered the minimum adjustments necessary to abide by WTO rules. The reform route soon proved unappealing due to a split among institutional Challengers (Hypothesis 2)—between countries like Chile and Australia who, as major wine exporters, depended on access to Europeans markets and, as a result, were relatively amenable to compromise, and countries like the UK (which exclusively imports wine) and the US (which exports only 15% of its wine production) who favoured far-reaching deregulation and were unwilling to align with the lower reservation point of fellow Challengers (wwtg.org; Silverman et al., 1999). By contrast, the OWC coalition was more united given that most European OIV members were also EU members and therefore subject to joint protocols and directives regulating the common European market in wine which served to closely align their regulatory interests (EU Council Regulation 479/2008; Meloni & Swinnen 2013).

Despite their greater unity, European wine producers were not the first to move. Instead, the first attempt at enacting institutional change took the form of an “indirect” rival creation. In 1998 seven NWCs (Argentina, Australia, Canada, Chile, New Zealand, South Africa, and the US) formed the World Wine Trade Group (WWTG) to promote “unsubsidized wine production and free export markets” and to “enhance the international acceptability of wine produced in WWTG countries” (wwtg.org). The group, which was soon joined by Georgia and Uruguay, accounted for about 1/3 of global wine production and exports. Although it didn’t amount to a direct rival to the OIV with conflicting regulations (loss of scale-economies made the co-existence of two rival regulatory regimes for wine unpalatable to all players, condition #1), by dedicating itself to lobbying the WTO to adopt policies favourable to NWC producers, and by specifically promoting “new world wine” (New Zeeland, 2001; wwtg.org), it amounted to a form of painful reneging whereby institutional Challengers inflicted losses on Defenders (condition #3).

As our theory predicts, given the high cost of ending Challenger’s reneging through reform (thanks to the outlying preferences of the US and UK), Defenders used their greater unity to instead launch a Succession. In June 2000, France convened a conference of 36 of the OIV’s 47 members—a large “core group”—to negotiate a new agreement, and on 3 April 2001, 35 countries signed a new treaty creating the International Organization of Vine & Wine (IOV) to replace the OIV. In a typical succession move, the IOV’s founding treaty made consent to terminate the 1924 OIV treaty a precondition for joining the new IOV (Resolution Comex 2/2002). As a concession to NWCs, voting power was redistributed in their favour, and consensus decision-making replaced majority rule on questions of production standards, thus granting NWC producers the power of veto (New Zeeland MFA, 2001). However, the IOV remained wedded to AOC standards “insofar as they do not call into question international agreements relating to trade and intellectual property” (Art.2.2.b.ii). Thus, defensive succession allowed OWCs to end painful reneging by offering more limited concessions to NWC than would have been required to achieve reform while continuing to defend the main privileges they had enjoyed as institutional leaders of the OIV.

By 2004, the necessary ratifications required for the IOV to enter into force were obtained. However, twelve OIV members declined to accept the new agreement, possibly gambling that it would not reach critical mass, or simply finding the price too steep. Many of these holdouts have since joined, but a small group have remained aloof—most notably Canada, the UK, and the US, whose initially higher reservation point made reform a more costly option to Defenders than succession. Despite the “fait accompli” of succession, the US and the UK have chosen to remain outside the new IOV. Our explanation is that Defenders made only enough concessions to attract Challengers who–like Chile or Australia–prioritized access to European markets. However, in 2006 the US and the European Community signed a bilateral agreement in which they recognized each other’s winemaking practices and exempted each other’s products from certification requirements. With this deal, the US (re)gained access to European wine markets in exchange for respecting certain “names of origin” such as Champagne for wines originating in the EC. This eventual accommodation shows what institutional succession can achieve subject to scale-economies and a united “core group”: bypass hard-nosed veto players and transform opposition into support.

3.2 The succession of the international sugar council (1968)

The International Sugar Council (ISC) was created in 1937 to stabilize trade in sugar for an initial period of five years, subject to review after that. The agreement, which instituted strict import/export quotas to stabilize prices, divided institutional power among two coalitions: sugar importing counties (UK, US, USSR, Holland) and sugar exporters (the remaining 18 parties). The Council was dominated by the two largest importers, the US and the UK, who together controlled 76% of importers’ votes and 38% of total votes, giving them de facto control of the agenda.Footnote 16 The institutional power of the UK and the US was further boosted by bilateral trade deals with individual sugar exporting countries (the UK with Commonwealth producers; the US with Cuba and other Latin American producersFootnote 17) which meant that direct imports from these countries to the UK/US were not counted as part of ISC import/export quotas (Swerling, 1954; Viton, 2004).

World War II delayed the treaty review scheduled for 1942. Instead, a wartime protocol was added which prolonged the existing agreement until 1944. At the same time, the war dealt a significant asymmetric shock to the ISC by prompting a change in Britain’s colonial policy (Viton, 2004:6) (condition #2). To stimulate economic development in sugar-producing Commonwealth countries where low prices kept workers in poverty, London announced a plan to replace British imports under the ISC with greater imports from the Commonwealth through a new Commonwealth Sugar Agreement (Moynagh, 1977; Mahler, 1984). Non-Commonwealth producers, Cuba especially, objected vehemently, since this would mean reduced British demand for their exports (Hull Telegram, Jan. 1944; Winant Telegram, March 1944). Siding with Cuba, Washington strongly pressured Britain to maintain the existing agreement, but agreed to temporarily suspend its operative clauses, thereby leaving Britain free to increase imports from the Commonwealth without violating the treaty (Telegram Hull, July 1944; Art.2, ISA Protocol, 1944). Cuba in turn responded to these degraded conditions by increasing its sugar production far beyond the existing ISC quota and unloading the excess onto world markets, causing world prices to drop by as much as 30% over the following decade (Mahler, 1984:716; IBRD, 1968; Viton, 2004:270)—a clear case of harmful reneging (condition #3).

Given that falling world prices harmed all exporters, Commonwealth producers included, Cuban reneging increased pressure on London to resume cooperation under the ISC. However, the pressure was insufficient to bring the parties to the table to commence a negotiation of quotas. Deadlock reflected lack of unity within both coalitions. Firstly, the split between the UK (who promoted the welfare of Commonwealth producers) and the US (who supported Cuba) hindered reform, as the two leading importers were incapable of concerted action. Second, exporting countries were also split given that Commonwealth producers were relatively less harmed by falling world prices than other exporters (since they were partly compensated by higher prices paid by the UK for non-ISC regulated imports). With neither side sufficiently united to drive through reform or initiate a succession, the new status quo prevailed (Hypothesis 3).

Given the infeasibility of either reform or a defensive succession (on account of the split between the UK and the US), institutional change was eventually achieved through an offensive succession. Two variables—the degree of harm caused by reneging and the unity among exporting countries (the institutional Challengers of this story)—paved the way for this move:

First, Castro’s ascent to power in Cuba in 1959 dealt a second unexpected shock to the ISC by leading President Eisenhower to redistribute Cuba’s US export quota to other Western Hemisphere countries (Berman & Heinemann, 1963; IBRD, 1968). Facing a sudden US embargo, Havana instead turned to Moscow who agreed to buy the bulk of Cuba’s sugar production and re-export it to the global market. This increased reneging caused world prices to drop to their lowest point since World War I.Footnote 18 Second, the creation of UNCTAD in 1964, where a stronger developing country vote existed, boosted the capacity of sugar exporting countries to organize outside the ISC. Between 1965 and 1968, some fourteen meetings were held at the UNCTAD to negotiate a new institution to regulate international sugar trade (Fakhri, 2014:186). In December 1968, an agreement was signed establishing a new International Sugar Organization (ISO) to replace the ISC. The new ISO treaty was signed by a larger number of developing countries than the incumbent ISC. It stated that the ISO would replace the ISC once ratified by countries controlling 60% of exporters’ and 50% of importers’ votes in the ISC (as opposed to the 75% majority required to reform the ISC). By lowering the threshold for acceptance, the ISO treaty thus effectively removed developed sugar importers’ veto and presented them with a fait accompli: either accept the ISO or leave the regime altogether.

As our theory predicts, given that it arose from an offensive succession, the ISO favoured institutional Challengers. Export restrictions once again became the chief device for stabilizing global prices (Art.28,1968-ISA; IBRD, 1968). The new agreement also entailed measures to broaden access to developed country markets (Art.51) and introduced a special fund to stimulate economic development in low-income sugar-producing countries (Art.1). Despite initial opposition, EC countries eventually joined the ISO as a bloc. The UK also joined, receiving a significantly smaller vote share (15%) than under the ISC. The ISO eventually included 46 exporting and 30 importing state parties with votes split 50/50 between the two groups. Among original ISC members only the US refused to join, accusing the new institution of favouring Cuban and Soviet interests. In a rare feat, Cuba and its supporters, thanks to their greater unity, were able to assume control of the international sugar institution by means of an offensive succession (Hypothesis 2).

3.3 The succession of the GATT

The Uruguay Round (1986–1994) which established the WTO as a successor to the GATT is another case of defensive succession to bypass outlying veto players. The GATT was revised through eight rounds of multilateral trade negotiations between 1948 and 1995 before being replaced by the WTO. Why was institutional change achieved through reform until the 1980s, but through succession in 1995? The answer, we argue, lies in the growing cost of corralling support for new reforms. We are not the first to argue that the WTO presented a way to break stalemate in the GATT (e.g., Steinberg, 2002; Jupille et al., 2013; McKibben, 2015). Unlike existing accounts, however, our theory explains the transition from the GATT to the WTO as a standard solution to a deadlocked reform process, and accounts for why succession did not happen earlier.

Trade rounds fundamentally are package deals, whereby a decision is taken simultaneously on different issues to achieve a balance of gains and concessions. For linkage to be manageable, states typically need to form coalitions on specific issues and then trade these issues as blocs (Hoekman, 1989:705). In the GATT, developing countries have often negotiated as one bloc, and the EC, the US and Japan as another (Hoekman, 1989:701; Steinberg, 2002).

The EC and the US have dominated bargaining and outcomes at the GATT from its early years, despite adherence to consensus-based decision-making (Steinberg, 2002:340). However, their institutional leadership has not been unchallenged. From the 1960s, the power of developing countries within GATT began to grow (Hudec, 1992). First, decolonization increased the number of developing country members so that by 1970, 52 of 77 GATT parties (68 percent) were LDCs (68 percent)—rising to 77 percent in 1991. Developed countries retained agenda-control by moving from voting to decision-making by consensus, thus preventing a supermajority of LDCs from controlling legislative functions of the GATT (Steinberg, 2002). Nevertheless, for developing country “institutional followers” there was power in numbers.

Second, developing countries became better organized politically. In 1964, they created the UN Conference on Trade & Development (UNCTAD) which led to the formation of the G77. UNCTAD was never a viable rival to the GATT. With the US and the EC accounting for two-thirds of the global economy, and most G77 countries trading heavily with developed countries rather than with each other, the idea of LDCs breaking away from the GATT was impracticable (scale-economies foreclosed rival creation, condition #1). However, UNCTAD served to galvanize the G77 coalition, allowing LDCs to win several concessions from developed countries (Jupille et al., 2013:84). For example, in 1964, the GATT parties added Part IV on “Trade & Development” to create special trade opportunities for developing countries, and in 1968 they added the Generalized System of Preferences (GSP) whereby developed countries could offer preferential tariff treatment to developing countries on a voluntary basis (Hudec, 1992).

Package deals are generally easier to conclude than issue-by-issue negotiations, as negotiators can more easily make necessary concessions by linking them to tangible benefits across issues (Hoekman, 1989:695). But for such linkages to be feasible, the quid and the quo must be comparable. Such was the case in the early trade rounds within the GATT, when negotiations mainly bore on easily quantifiable tariffs. However, negotiations got trickier when members turned to liberalizing non-tariff barriers (NTBs) during the Tokyo Round (1973–1979). This meant harmonizing domestic regulations in areas as diverse as customs valuation, import licensing procedures, anti-dumping practices, packaging and labelling, and public procurement. As a result, many agreements concluded during the Tokyo Round were “plurilateral” stand-alone agreements (or so-called NTB codes) to which only some states adhered. Because the codes were negotiated primarily by developed countries, few LDCs signed on, but LDCs were nevertheless able to gain the benefits of the codes through MFN treatment—a concession they secured by threatening to block the Tokyo Round agreement (Jupile et al., 2013:89–90). Reform was achieved at the expense of coherence (Krasner, 1979).

Reform was achieved in the Tokyo Round because both coalitions—developed country Defenders and LDC Challengers—were unified (Hypothesis 1). However, the fact that developing countries received the benefits of NTB codes without accepting the associated obligations led Defenders to complain of a “free ride” (Steinberg, 2002:357). This raises the question: Why did Defenders concede reform rather than pursue change through (temporary) regime-shifting or succession? Jupille et al. (2013:147–8) argue that developed counties could have liberalized NTBs at lower cost by shifting the issue to the OECD (whose members included a majority of the signatories of the NTB codes) but stuck with the GATT due to bounded rationality. In fact, rather than pursuing a strategy of inclusive reform myopically, Defenders considered several alternative strategies—as synoptically rational players do. During the Tokyo Round, US negotiators floated the idea of breaking LDCs’ institutional lock on the reform process by creating an alternative preferential regime—a “GATT-Plus”—through which the US, EC, and other industrialized countries would deepen trade liberalization among themselves, extending the benefits only to countries willing to accept the obligations.Footnote 19 The result would have been a two-tiered global trade regime, which would gradually pressure more LDCs into lowering NTBs to avoid trade diversion (Hufbauer, 1989; Steinberg, 2002:359). However, the strategy was dropped as heightened Cold War tensions dampened US willingness to coerce developing states (Steinberg, 2002:366).

The compromise reached at Tokyo was short-lived. Soon after the end of the Round in 1979, the US began to press for a new negotiation to address old issues left over from Tokyo—e.g., agricultural trade—and to expand the GATT to include new areas like services, investment, and intellectual property. The timing of US demands for renewed reform merits brief explanation. On our theory, institutional change occurs in response to some change in the environment that leads to an unexpected shift in relations between parties to an agreement (condition #2). As Winham (1989) observes, to some extent, no exogenous shift is needed to trigger negotiations within the GATT/WTO, since the idea of negotiation rounds is built into the regime. However, American pressure to re-open reform discussions was sparked by several environmental changes which led the US to wish to revisit the deal struck at Tokyo. The US economy—along with most European economies—was becoming increasingly dependent on the production of services such as banking and telecommunications that were not covered by the GATT. Pressure for reform was further reinforced by a growing US trade deficit which led the Reagan Administration to push for a renegotiation of the GATT to pre-empt protectionist demands at home (Winham, 1989).

Leading up to the Uruguay Round states agreed an all-encompassing agenda,Footnote 20 covering virtually all conventional trade issues plus new trade-related issues like intellectual property and foreign investment—two issues that pit OECD countries (and especially the so-called Quad, including Canada, EU, US, and Japan) against developing countries. Yet neither coalition was united. Transatlantic differences over agriculture split the US from the EC as key European countries refused to accept the G-40 text on agriculture (Winham, 1989:295). The US, the EC, and Japan also took separate positions on safeguards (Hoekman, 1989). The call to bring services into the GATT was strongly resisted by a group of LDCs led by India, Brazil and South Africa who argued they could not negotiate services on an equal footing with developed countries (Jupille et al., 2013:157). One of the legal tactics they pursued was to define services as outside the competence of the GATT. As a result, the special session in Punta del Este in September 1986 which launched the Uruguay Round created an entirely separate negotiating structure for services, thereby reducing the prospects that institutional Defenders would demand concessions from Challengers in services in return for concessions on goods (Winham, 1989:291). Negotiations were deadlocked for several years, but by 1994 separate “negotiation groups” had reached separate agreements on trade in goods (GATT, 1994), trade in services (General Agreement on Trade in Services, GATS), Trade-Related Intellectual Property Rights (TRIPs) and Trade-Related Investment Measures (TRIMs). Still, an overarching agreement to close the Round proved elusive as many developing countries stated their intention not to sign the agreements on TRIPs, TRIMs, or GATS (Steinberg, 2001:359) reducing the value of reform for Defenders.

From our theoretical perspective, the GATT negotiations in the early 1990s were at a similar stage as the international wine office in the 1990s: a small core of unified Defenders were willing to grant limited concessions to Challengers but saw the road to agreement blocked by a handful of powerful “outlying” Challengers who demanded greater concessions than Defenders were willing to give. In both cases, this meant that Succession was the preferred route to institutional change (Hypothesis 2).

Realizing that the reform route would mean that many countries would opt out of what they considered to be key bargains, the Quad (acting as a core group) abandoned the stand-alone agreements format of the Tokyo Round and opted for a succession. To do so, they included all agreements reached in Uruguay—the GATS, TRIPs, and TRIMs agreements; the Subsidies Agreement; the Anti-dumping Agreement plus every existing agreement concluded under GATT auspices—in one package deal, which, for countries that signed it, terminated all extant obligations under the 1947 GATT Agreement. After joining the WTO, the Quad withdrew from the GATT of 1947, thereby terminating their existing GATT obligations (including the MFN guarantee) to countries that did not accept the Uruguay Final Act (Steinberg, 2001:360). As Hudec (1992:76) notes, “in essence this…completely restructured the developed-developing countries bargain, proposing to pay for all the new developing country concessions simply by agreeing not to destroy the market access they already [had].” This move, appropriately named the “single undertaking’, enacted a succession through which institutional Challengers were forced to accept all aspects of the Uruguay Final Act or be excluded from the global trade regime, since the GATT was effectively abolished (Patterson & Patterson, 1994; Steinberg, 2002:360). As a result, many countries ended up accepting measures such as the TRIPS and TRIMS agreements, which they would have blocked if debated on the plenary floor or refused to ratify if written as stand-alone agreements. The decentralized mode of consent that is typical of succession allowed the Quad countries to recreate the original coherence that had been lost in the Tokyo Round, while pushing the content of the trade agreement beyond its former limits.

3.4 IMF reform

Formally provided for by most IGO constituent treaties, reform is the most inclusive process of institutional change. It is also difficult to achieve since it tolerates no or little dissent. The previous case studies have featured instances of both stasis, reform and succession depending on the relative unity of opposing coalitions. Here we further illustrate our theoretical propositions by briefly analysing the reform of the IMF.

The IMF was founded in 1944 to promote exchange rate stability and monetary cooperation through the provision of short-term funds to countries facing balance of payments difficulties. Throughout its history, the Fund’s Articles of Agreement have been amended seven times (Bordo & James, 2000). Only two of these reforms have been controversial: the second amendment (adopted in 1973) and the sixth and seventh amendments (adopted in 2010) on which we focus here.

Voting power in the IMF long reflected founding members’ initial quotas, in turn a reflection of their relative economic wealth in 1944. Yet, as time passed, the growth of emerging market economies necessitated a reallocation of IMF financial quotas (and associated votes) from the smaller European founding members and Canada to mainly the BRICS in order to keep the Fund’s resources adequate to members’ financing needs. For many decades, China, which obtained a single-country seat on the IMF Executive Board in 1980, has used this position to argue for increasing the power and influence of emerging market economies (Momani, 2015:267). Calls for reform grew more intense in the wake of the Asian financial crisis and the Global Financial Crisis in 2008 (ibid). However, so long as European founding members (the institutional Defenders of this case) did not need IMF loans—they had not done so since the 1976 UK sterling crisis—they used their veto powerFootnote 21 to block any change that might weaken their grip on the institution (Vestergaard & Wade, 2014). With the path to reform blocked institutional Challengers had little choice but to respect the status quo since they lacked the financial muscle to initiate an offensive succession and had little incentive to engage in rival regime-creation (unlike multilateral development banks, the IMF’s efficacy rests on its singularity, condition #1).

This situation changed dramatically with the unexpected and asymmetrical shock of the 2010 Eurozone crisis (condition #2) which forced Ireland, Cyprus, Spain, and Portugal to borrow from the IMF, and Romania, Hungary, and Latvia to sign standby agreements. Asked to agree to a doubling of IMF resources, emerging markets conditioned their agreement on institutional Defenders’ consent to reform existing IMF quotas and voting weights. Under financial pressure, European IMF members withdrew their veto, and the IMF Executive Board agreed to redistribute 6.2% of their quota shares mostly to China, India, Brazil, Mexico, and South Korea (6th amendment), and by ending the practice of ‘appointed’ executive directors which had traditionally secured a permanent seat for the top five contributors (7th Amendment) (Diaz & Saz-Carranza 2015). Adopted in 2010, both amendments became effective in 2015 (IMF 2015).

The outcome of this case is easily explained by our theory: Until 2010 the IMF was ‘static’ due to European opposition to a financial quota/vote redistribution, and due to the inability of Challengers to either increase pressure for reform through reneging or mount an offensive succession (Hypothesis 3). The effect of the 2010 Eurozone crisis was to unite institutional Defenders behind the realization that recapitalization had become a financial necessity, leading European members to rescind their veto over a reform that China and other emerging market economies had demanded for decades.

4 Limitations

Our theory of institutional change offers a stylized explanation for processes of reform versus succession. As is always the case, reality is less neat than the theoretical models we build to explain it. Nevertheless, our theory captures the main parameters leading institutions to either remain static or to undergo change through either reform or succession.

Although it offers a comprehensive theory of institutional change, our theory does not cover all scenarios. First, positing a tipping dynamic (premised on scale-economies) as a scope-condition means out theory cannot predict the establishment of rival IGOs. Rival creations frequently occur (see Morse & Keohane, 2014: Urpelainen & van de Graaf, 2014), but only, we argue, in cases where the loss of scale-economies is not prohibitive (which lie outside the purview of our theory). An illustrative example is the Ottawa Convention established in 1999 as a rival to the Convention on Certain Conventional Weapons (CCW) which regulated the use of anti-personal landmines since 1980. As negotiations over more stringent standards grew deadlocked, dissatisfied states chose to pursue a new agreement in parallel to the CCW (Cottrell, 2009; Morse & Keohane, 2014). In this case, the absence of significant scale-economies implied that states willing to adopt more stringent standards did not need to convince all other relevant players to join them in order for their efforts to yield substantive benefits. The parallel Ottawa Convention did not lead to a loss of utility through unenforced cooperation within the existing CCW regime.

5 Conclusion

This article introduced the concept of institutional succession as a regular mechanism of institutional change in multilateral cooperation and has theorized succession in relation to other strategies of institutional change. Working from a definition of succession as the creation of a new IGO that formally assumes all (or the major part) of an existing IGO’s mandate and functions, we articulated the pros and cons of succession relative to other forms of institutional change. We proceeded by theorizing a bargaining situation that explains when one of three alternative outcomes can be expected: (i) continued use of an existing institution that has been rendered inefficient by changing environmental circumstances (‘stasis’); (ii) change through reform, or (ii) institutional replacement through a succession that is either controlled by institutional Defenders (defensive succession) or initiated by Challengers (offensive succession).

Contrary to the literature’s belief that succession is a costly move of last resort, we have demonstrated that reform and succession are equally plausible and appealing mechanisms of institutional change but under different circumstances. On the one hand, reform requires that both negotiating sides be sufficiently unified to overcome veto players. In contrast, succession only requires a subset of members from either side to propose a successor institution which a minimum threshold of other parties can be persuaded to join. On the other hand, succession suffers from scale diseconomies; not every party to an incumbent organization may join the successor, thus reducing overall efficiency gains compared to reform, which tends to involve all existing parties. Depending on which shortcoming prevails, we predict that reform will be preferred to succession or vice versa. Our empirical findings bear this out: in bargaining conflicts that feature split coalitions—as was the case in the two commodity agreements we analysed as well as the GATT—succession prevails over reform. In contrast, in situations where both sides have been cohesive, as in the IMF (and the GATT’s Tokyo Round), reform has prevailed.

By theorizing an overlooked phenomenon—the creation of new IGOs to replace existing ones—and by anchoring our theory of institutional succession within existing literature on contested multilateralism, we seek to open new avenues for research on mechanisms of institutional change. Our analysis suggests two immediate avenues for future research. First, our theory allowed us to account for key changes in the histories of four IGOs regulating trade in wine (OIV), in sugar (ISO), general global trade (GATT), and international finance (IMF), and to account for the different evolution of institutions across the domains. This provides a strong indication of the empirical plausibility of our theory. Nevertheless, more empirical cases are needed to further increase confidence in our theory and to identify potential omitted variables. Second, future extensions of the theory may benefit from relaxing certain simplifying assumptions (such as the two-coalition bargaining set-up) and from theorizing factors not directly considered here such as the choice of regime-shifting v. other inside options, or factors facilitating gradual and informal institutional change (which may omit the need for either reform or succession).