Analogy-based collective decision-making and incremental change in international organizations

Abstract

We examine how analogy-based collective decision-making of member states contributes to the endogenous emergence of informal rules and the incremental change of international organizations (IOs). Decision-making by analogy is an important characteristic of day-to-day decision-making in IOs. Relating current decisions to previous ones through analogies drives incremental change and simultaneously reinforces organizational resilience. Whereas the foreign policy analysis literature shows that analogies can be used as cognitive shortcuts in fuzzy and complex foreign policy situations, we focus on their use to overcome social ambiguity (indeterminacy) of coordination situations in IOs. Drawing on psychological conceptions, we develop two micro-level mechanisms that elucidate the effects of analogy-based collective decision-making in member-driven IOs. Analogy-based collective decisions emphasizing similarity between a current situation and previous ones follow an established problem schema and produce expansive and increasingly well-established informal rules. Collective decisions that are analogy-based but emphasize a crucial difference follow different problem schemas and trigger the emergence of additional informal rules that apply to new classes of cases. The result is an increasingly fine-grained web of distinct organizational solutions for a growing number of problems. Accordingly, an IO can increasingly facilitate collective decision-making and gains resilience. Empirically, we probe these propositions with a documentary analysis of decision-making in the Yugoslavia sanctions committee, established by the United Nations Security Council to deal with a stream of requests for exempting certain goods or services from the comprehensive economic embargo imposed on Yugoslavia in response to the War in the Balkans.

Introduction

We examine how analogy-based collective decision-making of member states contributes to the endogenous emergence of informal rules and the incremental change of international organizations (IOs). We focus on member state-driven organizations without a powerful bureaucracy, such as the United Nations (UN) Security Council. We understand such IOs as sets of rules and routines that provide decision criteria and facilitate horizontal day-to-day decision-making of member states. In organizations, “rules and situations are related by criteria of similarity or difference and through reasoning by analogy and metaphor” (March and Olsen, 1989: 25). Analogy-based collective decision-making helps overcome the social ambiguity in coordination situations in which member states need to agree on one of several solutions. It rests on the relational similarity of a current situation with previous ones and creates path dependence (David, 1994). Like analogy-based judicial decision-making (Fon et al., 2005; Stone Sweet, 2004), which produces legal doctrines, it may tacitly produce organizational rules and routines that change incrementally if they are applied to ever new cases. Hence, analogy-based day-to-day decision-making endogenously drives incremental change and simultaneously reinforces organizational resilience.

Current International Relations theories are unsuited to analyzing the role and effects of analogy-based collective decision-making in day-to-day operations as an endogenous source of incremental change and organizational resilience. Rational institutionalism focuses mainly on formal procedures and understands institutions largely as exogenous constraints and “time-invariant rule of the game” (Fioretos, 2011: 372). It perceives international institutions as inherently stable cooperation arrangements (Keohane, 1984) or a series of grand bargains (Moravcsik, 1998), while the endogenous effects of day-to-day operations are largely ignored. This helps understand the design of IOs but is inadequate to explain how IOs operate and develop over time (Stone Sweet, 2004: 9). Constructivists attribute organizational dynamics mostly to autonomous bureaucracies or courts (Barnett and Finnemore, 2004), but largely ignore intergovernmental decision-making. While constructivist practice theory draws attention to powerful informal rules and practices in social life, including international institutions (Adler and Pouliot, 2011), it has difficulty explaining how strategic actors become enmeshed in social practices.

Historical institutionalism provides a fruitful starting point because it advocates a shift of perspective from institutional choice to institutional development (Pierson, 2004: 15). First, it takes temporality seriously and addresses the effects of sequential decision-making for organizational development. Fioretos (2011: 371) points out that “attention to the timing and sequence of political events is important because the evolution of constraints and opportunities in [. . .] institutions [. . .] often create a different type of political game over time.” In path-dependent processes, the causes that trigger establishing an arrangement typically differ from those that stabilize the arrangement once established (Mahoney, 2000: 512). Second, historical institutionalism pays attention to endogenous sources of change, while rationalist approaches typically “can only account for change to or away from an equilibrium, but cannot make sense of changes within an equilibrium” (Rixen and Viola, 2016: 14). Endogenous effects occur as feedback loops—that is, institutional effects become the cause of subsequent effects in the same system, which in turn become causes again (Rixen and Viola, 2015: 305). Informal rules evolving from analogy-based collective decision-making result from such endogenous processes. Third, historical institutionalism focuses on self-reinforcing processes. These processes produce increasing returns and occur in situations “in which the returns actors derive from following an institutional rule or practice increase relative to the initial investment” (Rixen and Viola, 2015: 305), such as when following an emerging informal rule benefits an increasing number of actors or facilitates collective decisions in ever more situations. IO scholars are just beginning to take temporal, endogenous, and self-reinforcing processes seriously (Fioretos, 2011; Rixen and Viola, 2016). Current analyses mostly perceive organizational path dependence as a pathology driven by a lock-in of outdated rules (Hanrieder, 2015; Zangl et al., 2016) or focus on the evolution of vested interests of societal groups that sustain institutional arrangements (Büthe, 2016). In contrast, mechanisms that shape IO decision-making, such as evolving informal rules through analogy-based collective decision-making, are still missing.

Drawing on psychological conceptions (Gentner and Smith, 2012; Novick, 1988), we develop two micro-level mechanisms that elucidate how analogy-based collective decision-making in member-driven IOs may create incremental change and organizational resilience. For this purpose, we focus on analogy-based decision-making to overcome social ambiguity in collective decision situations with multiple equilibria. This differs from other forms of analogical reasoning, such as providing cognitive shortcuts in fuzzy, ill-structured, and unique foreign policy situations, as emphasized by the foreign policy analysis (FPA) literature (Hemmer, 2000; Houghton, 2001; Khong, 1992). We also assume strategic action of member states. While constructivist branches exist, both historical institutionalism and analogical reasoning are compatible with strategic action. Pierson (2004: 9) stresses that “rational choice analysis, broadly defined, offers essential analytical tools for investigating temporal processes,” and Fioretos (2011: 376–377) emphasizes that the study of micro-level mechanisms helps avoid a deterministic understanding of institutional development. Likewise, the FPA literature demonstrates that analogical reasoning concepts can be firmly situated within a rationalist research agenda, assuming that actors employ analogies to identify suitable strategies to pursue their interests (Hemmer, 2000: 26).

The theoretical argument proceeds in five steps. We first explore the demand for analogy-based collective decision-making to overcome social ambiguity in indeterminate coordination situations. Next, we outline a multi-stage process of analogy-based collective decision-making that distinguishes between stages performed individually including invoking an analogy, and those performed collectively, especially evaluating an analogy and ultimately deciding. Subsequently, we develop two micro-level mechanisms that reveal how analogy-based collective decision-making produces both incremental change and organizational resilience. If actors emphasize the similarity of a situation with previous ones, they can follow the problem schema of previous decisions. This mechanism leads to self-reinforcing and expansive (“drifting”) informal rules that turn non-routine tasks into routine ones. If actors recognize the analogy to previous cases but reject the resulting solution, they trigger a reactive sequence that establishes a new category of cases and simultaneously an analog for future cases of this type. This mechanism leads to a layered set of informal rules that offers solutions for a growing number of problems. Finally, we discuss the consequences of analogy-based collective decision-making for IOs and the implications for its empirical analysis.

Empirically, we probe analogy-based collective decision-making in the Security Council’s Yugoslavia sanctions committee, which managed the Yugoslavia sanctions regime. Evaluating unique internal documents, we demonstrate that the committee relied heavily on analogies, that the use of analogies shaped collective decisions following our two mechanisms even when members widely disagreed over implementation, and that the Council over time offered an ever denser set of solutions for different classes of cases. The Security Council is ideal for studying evolving informal rules. First, member states, particularly the five powerful permanent members, dominate this “high politics” organization. Council members establish committees that mirror Council membership and decide by consensus to control decisions tightly and avoid delegation to the UN Secretariat. Second, the Yugoslavia sanctions committee had wide-ranging discretion over how to implement the sanctions regime. Its rules and routines were not designed ex-ante, but evolved over time. Third, the committee dealt with a stream of similar but non-identical requests exempting certain goods or services from the trade embargo imposed on Yugoslavia. Hence, rules and routines evolved within a short time frame, minimizing external shifts in interests. Insights from Security Council committee governance are likely to matter for many IOs, which process streams of non-identical decisions, such as listing endangered species under the Convention on International Trade in Endangered Species of Wild Flora and Fauna (CITES, allocating funds in the World Bank or regulating pharmaceuticals in the European Union (EU).

The careful analysis of analogy-based collective decision-making emphasizes the potential provided by a historical institutionalist analysis of how IOs operate and develop (Rixen and Viola, 2016: 6–8). First, we isolate analogy-based collective decision-making as a source of influence that may occur within a relatively short time frame. It shows that historical institutionalism is relevant for studying the endogenous effects of day-to-day decision-making in IOs (Fioretos, 2011: 373–376), while it has mainly been used to examine major societal or institutional developments over extensive time horizons (see International Relations chapters in Fioretos et al., 2016). Second, our concept accounts for both organizational resilience and incremental change of analogy-based rules and routines. Organizational resilience is reinforced both by analogy-based rules, which are expansive and increasingly established, and by additional layered analogy-based rules and routines. Change does not reflect “the movement to a different equilibrium – it does provide an explanation of change along a path – that is, the increasing entrenchment of an equilibrium” (Rixen and Viola, 2015: 310). Third, we contribute to the debate on the role and effects of IOs. Identifying an endogenous influence that originates exclusively from repeated interaction among member states, we complement concepts of IOs focusing on principal-agent arrangements (Hawkins et al., 2006), bureaucratic agents (Barnett and Finnemore, 2004), or institutional design (Koremenos, 2013).

Analogy-based collective decision-making in IOs

Analogical reasoning is the ability to perceive and use relational similarity between two situations (Gentner and Smith, 2012: 130). It rests on the basic premise that when two knowledge domains are significantly similar, the two can be treated as instances of the same problem. Information from the familiar “source” domain may be used to overcome information gaps about the unfamiliar “target” domain (Peterson, 1997: 248). As Kratochwil (1989: 223) noted, “the task of analogies is to establish similarities among different cases or objects in the face of (striking) dissimilarities.” Analogical reasoning is important for foreign policy-making by governments and parliaments in international crises (Hemmer, 2000; Kaarbo and Kenealy, 2017; Khong, 1992), in other foreign policy situations (Breuning, 2003; Vertzberger, 2002), and for precedent-based legal decision-making (Lamond, 2016; Schauer, 1987; Stone Sweet, 2004), which may be considered a “formalized application of analogical reasoning” (Holyoak, 2005: 117). As “a fundamental aspect of human cognition” (Gentner and Smith, 2012: 130), analogical reasoning is also highly relevant for day-to-day decision-making in IOs.

The demand for analogy-based collective decision-making in IOs

Drawing on psychological accounts (Gentner and Smith, 2012; Novick, 1988), analogies have acquired an important role in FPA. A rational branch of FPA perceives analogies as “cognitive tools” that may be “employed to make sense of a complex reality” (Houghton, 2001: 22). In major crises, such as the Cuban missile crisis (Tierney, 2007), the Vietnam War (Khong, 1992), the Iran hostage crisis (Houghton, 2001), and the 9/11 terrorist attacks (Mumford, 2015), key decision-makers such as US presidents have drawn on analogies with historical situations. In such unique foreign policy situations, decision-makers suffer from bounded rationality, while “lessons from history” provide a cognitive shortcut to grasp a complex current crisis (Hemmer, 2000: 3–12). Analogies may help define the situation, assess stakes, and develop strategies (Khong, 1992: 10), or establish moral constraints on strategic choices (Tierney, 2007). Likewise, they may lead to policy failure “because no two cases are exactly alike” (Kaarbo and Kenealy, 2017: 69). Accordingly, analogies with previous situations may constitute independent variables that may heavily influence political decisions even if stakes are high and situations are unique (Houghton, 2001: 22). However, they may also be used to justify political decisions adopted for other reasons (Mumford, 2015).

Analogy-based decision-making also plays a major role for day-to-day operations of IOs. Organizations provide their service best in familiar situations, where they can draw on a repertoire of solutions. In such situations, they can process tasks according to their sets of formal and informal rules, standard operating procedures, and routines. Even in routine situations, IO decision-making is complicated because decision problems are never identical. Schauer (1987: 577) emphasizes from a legal perspective:

No two events are exactly alike. [. . .] Were that required, nothing would be a precedent for anything else. We must therefore leave the realm of absolute identity. Once we do so, however, it is clear that the relevance of an earlier precedent depends upon how we characterize the facts arising in the earlier case.

Analogical reasoning constitutes a typical component of IO decision-making because it relates situations through criteria of similarity or difference (March and Olsen, 1989: 25). We target analogical reasoning related to these day-to-day decisions. This does not imply that IO member states or bureaucracies cannot employ analogies to deal with exceptional high-stakes situations, such as developing outer space law in the UN (Peterson, 1997), which confront decision-makers with challenges similar to decision-making in foreign policy crises.

We focus on one major collective decision problem of member-driven IOs, which originates from social ambiguity produced by multiple equilibria of coordination situations. States establish member-driven IOs, including the UN Security Council, as incomplete contracts to adopt secondary decisions within institutional frameworks (Abbott and Snidal, 1998). IO members frequently face situations in which they prefer agreement over non-coordination but advocate different solutions. These coordination situations are characterized by multiple equilibria and confront actors with the problem of agreeing on one solution from a range of options (Snidal, 1985: 931–936). Their indeterminacy creates social ambiguity, even if actors’ preferences and other relevant facts are well-known and cognitive shortcuts are not needed. Social ambiguity arises from the horizontal setting of collective decision-making and is less relevant for situations in which the power to make decisions rests with the head of state/government or a bureaucracy. Indeterminate coordination situations abound in IOs. Due to high quorums (often consensus or unanimity), IO decisions often require broad support. The more IO decisions that are to be made, the more important smooth decision-making is and, accordingly, the more important it becomes to overcome social ambiguity through reference to previous cases. Even Security Council members frequently have a common interest in smoothly operating sanctions regimes agreed upon by a broad Council majority, despite different preferences regarding the design and nature of such regimes.

Analogy-based collective decision-making helps overcome social ambiguity in horizontal decision situations and represents an ideal-type concept. It is based on the analogy between a current case and previous ones. Previous decisions lend themselves as a collectively recognized reference (focal point) in indeterminate situations. They suggest “to repeat the action that succeeded before if we have no strong reason to do otherwise” (Lewis, 1969: 36–37) and are compelling for strategic actors because they have been collectively acceptable before (Yee, 1997: 1025–1027). Past experience, tradition, and precedent are known to have a stabilizing effect (Schelling, 1960: 260; Snidal, 1985: 936). When considering the use of analogies with past cases, actors make a “point-to-point” calculation (Fioretos, 2011: 373) of the benefits of repeating a previous solution with the costs of identifying a new one and the risk of non-coordination (also Houghton, 2001: 28). Both the coordination problem and the value of using the analogy increase with the number of actors and strategies (Snidal, 1985: 935).

Analogy-based collective decision-making to overcome social ambiguity differs from several other ideal-type forms of decision-making. In contrast to analogy-based cognitive shortcuts, which help overcome cognitive uncertainty (Hemmer, 2000: 5; Houghton, 2001: 21–23; Mumford, 2015: 6), it is relevant even if decision-makers are familiar with all the options and their respective implications. In contrast to bargaining, resulting decisions are based on reasoning, not merely on cost–benefit analysis, and do not necessarily reflect the preferences of powerful members. In contrast to other forms of reasoning, occasionally labelled “creative” (Houghton, 1996: 525) or “abstract” reasoning (Breuning, 2003: 232–233), which are based on the merits of a case and causal explanations, it refers to arguments regarding the analogy with previous cases. In contrast to reasoning processes that result in decisions by individuals, it results in collective decisions by IO members. In real-world situations, several of these forms may occur in combination.

Repeated analogy-based collective decision-making to overcome social ambiguity promotes the gradual evolution of problem schemas and related rules that reflect organizational solutions for a given type of problem. Houghton (2001: 26) emphasizes that analogical reasoning plays a key role in psychological concepts of schema formation (also Hehir, 2006: 72). When an actor has solved a problem successfully on two or more occasions, the actor will derive a “problem schema,” a set of abstract principles for dealing with this type of problem (Gick and Holyoak, 1983: 32; Khong, 1992: 25–26). While constructing schemas involves the use of analogy and metaphor, once formed, these schemas go beyond the instances that inspired them (D’Andracle, 1989: 810). The tacit emergence of informal rules from repeated interaction is also known from the rational choice concept of conventions (Lewis, 1969) and the constructivist concept of social practices (Adler and Pouliot, 2011; Gehring and Dörfler, 2019).

As a result, analogy-based collective decision-making to overcome social ambiguity matters in two different ways for day-to-day IO operations. First, it allows a solution to be found for a current problem through “reasoning from analog” (Gick and Holyoak, 1983: 9). Thereby, it triggers tacitly emerging rule-like schemas and gradually transforms non-routine problem-solving into routine problem-solving. Organizations are particularly strong in routine problem-solving precisely because they allow “reasoning from schema,” which makes salient those aspects of a situation that trigger a particular plan of action (Gick and Holyoak, 1983: 10). According to a standard distinction, routine problems are those for which schemas exist, whereas solving a non-routine problem requires a complex search and comprehension process (VanLehn, 1989: 545–550; also Houghton, 2001: 28). Second, analogy-based collective decision-making supports choosing an appropriate schema or rule in socially ambiguous situations, when either more than one schema is available or no schema covers the entire problem (VanLehn, 1989: 549–550). Psychologists acknowledge that actors frequently choose between different source analogs (Holyoak, 2005: 123) and FPA scholars highlight the decisiveness of such choices (Hemmer, 2000: 17–23; Houghton, 2001: 29–31).

The process of analogy-based collective decision-making

To grasp the complex process of analogical reasoning, we follow psychological concepts that have informed political science analyses. Ideally, this process has five steps (Gentner and Smith, 2012): (1) Representation: Actors develop a preliminary characterization of key features of the target domain. Different representations can lead to the use of different analogies and produce different conclusions about the target domain, as FPA scholars emphasize. (2) Retrieval: Actors identify possible analogs that show how similar problems have been dealt with before. This stage involves the ability to recall suitable analogs from individual or institutional memory, a process that may be distorted by cognitive constraints or interests (Hemmer, 2000: 17–21; Khong, 1992: 31–35). (3) Mapping: Actors assess how two situations are similar, and then bring across further inferences from the source to the target. Hence, this core process addresses the similarity or difference between the current decision and its analog (Gick and Holyoak, 1983; Holyoak and Thagard, 1996: 24–31). (4) Evaluation: After actors have aligned common features and generated inferences, they evaluate them (Gentner and Smith, 2012: 133). In IO decision-making, this stage includes recognizing the analogy as suitable with the consequence of following it or emphasizing dissimilarities with the consequence of not following it. (5) Generalization: Actors induce a problem schema from structural similarities of the source and target domains that covers a larger set of cases (Keane et al., 1994: 388–389).

The first two stages, representation and retrieval, are a matter of individual action. All actors, including neutral ones (e.g. the chairperson), may invoke an analogy. When they do so, they highlight certain characteristics of the current case and their similarity with those of a suitable previous case, while simultaneously disregarding differences. Actors can use analogies strategically (Kaarbo and Kenealy, 2017: 69–70; Sandholtz, 2008) by invoking analogies that support their interests (Gerhardt, 2005: 967–969; Hathaway, 2001: 135). Invoking an analogy may trigger a collective process of analogical reasoning, but it does not necessarily lead to analogy-based collective decision-making.

Stages three to five, mapping, evaluation, and generalization, are a matter of collective activity. To grasp this process in light of diverging preferences, we group IO members into three camps: the proponents of an invoked analogy, their opponents, and a group of members indifferent about the case. While an analogy might not persuade firm opponents, it can convince indifferent member states without stakes in the case (Krebs and Jackson, 2007: 42–48). If indifferent actors accept the analogy, they will support a corresponding decision. This decision promises to overcome the social ambiguity of horizontal coordination and does not infringe upon their interests. Accordingly, the camp of advocates of the analogy-based solution grows. Opponents can no longer expect a collective decision reflecting their preference and are pushed toward accepting the analogy-based solution, as long as they prefer agreement to non-coordination. Hence, analogy-based collective decision-making does not assume that all actors are convinced that the analogy-based solution is appropriate. It suffices that many of them are and that reluctant members give in to coordinate in a socially ambiguous situation. Such bandwagoning effects of evolving consensus have been widely observed for IOs, including the EU (McKibben and Western, 2014) and the Security Council (Monteleone, 2015: 51; O’Neill, 1996: 221). Analogy-based collective decision-making reflects agreement among the members that the current problem is sufficiently similar to previous ones to be solved along the same lines. Eventually, the underlying problem schema may be generalized to include a broader class of similar cases.

Identifying suitable analogies to overcome social ambiguity within day-to-day collective decision-making of IOs is less demanding than in analogy-based foreign policy-making. Actors deal with social ambiguity in typically well-known cases, not with “fuzzy, ill-structured tasks” (Houghton, 2001: 22–23). Decisions are less complex and simultaneously much denser, which increases the likelihood of finding suitable analogies. Most importantly, collectively compelling analogs are likely in short supply (Krebs and Jackson, 2007: 44–48; Stone Sweet and Sandholtz, 2004: 259). Hence, the criteria are not under the control of the individual members and may be assumed to be relatively stable over the short term (Schauer, 1987: 576–588). The fact that a broad range of actors with different interests collectively examines an invoked analogy tends to filter out analogs with an only superficial similarity between source and target domains (also Hemmer, 2000: 23–26). After all, what is at stake is not the ability to justify one’s own position but to find an analogy that actors with different preferences will commonly accept.

In the following, we identify emphasizing similarity and emphasizing difference (reactive sequence) as two analogy-based mechanisms of organizational development. We assume strategic action to ensure that emerging problem schemas and their effects are not attributed to internalization or habit. We also assume fixed interests, because we are interested in analyzing how IO decision-making drives endogenous change.

Emphasizing similarity: toward self-reinforcing and expanding informal rules

Whenever a group of actors adopts an analogy-based decision, it links two unrelated cases according to the same problem schema and triggers a gradually emerging informal rule. As a prerequisite for using the previously agreed solution as an analogy for the current problem, actors must collectively emphasize the similarity between the two non-identical cases according to key features. Accordingly, they create a class of cases that are considered alike in relevant aspects and can therefore be treated alike, despite striking differences in other respects. The repeated use of an analogy creates a generalized problem schema enshrined in an informal rule (Schauer, 1991: 181–187), which indicates how decisions of a given type are dealt with. The related rule emerges “spontaneously” from case-specific decision-making without formal collective agreement (Sugden, 1989). It may persist for a long time because it provides a common reference for members to make decisions in this type of situation (Pierson, 2004: 59). As VanLehn (1989: 545) notes: “If one gives subjects the same set of problems many times, they may learn how to solve them and cease to labor through [. . .] search processes.” The informal rule will become increasingly established with every new case decided according to the problem schema. It makes deviant behavior more difficult (though does not exclude it) because it puts an argumentative burden on deviators (Schauer, 1987: 581). The informal rule generates legitimacy within an IO because it reflects an accepted way of dealing with a type of problem, not because it is wise, fair, or problem-adequate.

The first decision determines the underlying problem schema. Hence, decisions are sequenced, earlier decisions are more important than later ones, and certain choices become more likely in later rounds because of preceding decisions (Rixen and Viola, 2016: 13). Generally, rules may be permissive, prescriptive, or proscriptive (Crawford and Ostrom, 1995: 584–585). Using a permissive decision as an analogy requires a permissive follow-up decision and precludes a proscriptive or prescriptive one. Otherwise, actors would leave the realm of analogy and resort to a different logic of decision-making. Hence, a permissive decision is likely to trigger a series of further permissive decisions, while a proscriptive or prescriptive decision is likely to trigger a series of further proscriptive or prescriptive decisions, as long as actors rely on analogical reasoning.

An analogy-based informal rule is expansive because its problem schema is applied (“drifts”) to an increasing number of non-identical cases. Every non-identical case decided according to the analogy expands the problem schema into a somewhat different realm. Slightly different cases are known as the key driver for evolving legal doctrines (Fon et al., 2005: 43–56; Pelc, 2014; Stone Sweet, 2004). They incentivize decision-makers to consider them as sufficiently similar to previous ones in order to facilitate collective decision-making, because rejecting the analogy might produce a stalemate. Strategic actors can select initiatives that have a chance of succeeding or generating a desired decision (Busch, 2007), because analogy-based decision-making increases predictability (Hathaway, 2001: 127–132). Actors may deliberately submit proposals that are slightly different from previous ones with the intention of broadening the problem schema. Every new decision allows it to be aligned with additional cases and paves the way for expanding the schema’s applicability. Accordingly, an innocuous decision may create a “slippery slope” and trigger similar but increasingly pernicious events (Schauer, 1985: 361–362):

Imagine a faculty meeting considering a request from a student for an excused absence from an examination in order to attend the funeral of his sister. [. . .] This case will establish a precedent allowing students to be excused from examinations to attend the funerals of grandparents, aunts, uncles, cousins, nieces, nephews, close friend, or pet. (Schauer, 1987: 578)

Such institutional drift is well-known in historical institutionalism. Drift “occurs when rules remain formally the same but their impact changes as a result of shifts in external conditions” (Mahoney and Thelen, 2010: 17). Whereas drift is often associated with stable formal arrangements in light of some demand for adaptation, in our case, actors choose to broaden the applicability of problem schemas to facilitate decision-making in non-identical situations.

An endogenously emerging analogy-based informal rule is self-reinforcing and generates increasing returns (Fioretos, 2011; Pierson, 2004: 17–53). Every additional case decided according to its logic expands the range of applicable situations. Given that following an established solution facilitates collective decision-making, the informal rule is increasingly useful for decision-makers because of this expansive applicability. Within the logic of analogy-based collective decision-making, the expansion of informal rules is difficult to contain. Actors may declare a decision as exceptional or not constituting precedent. Alternatively, they may collectively attempt to limit expansionary effects by defining a narrow scope for the problem schema, as legal analysts suggest (Schauer, 1987: 580). However, such declarations merely commit decision-makers themselves and may be ignored in later situations if deemed appropriate. They cannot diminish the incentive to base collective decisions on analogies with previous cases to overcome social ambiguity.

Emphasizing difference and reactive sequence: toward additional rules and increasingly fine-grained distinctions

Analogy-based collective decision-making may also result in what is known in historical institutionalism as a “reactive sequence” (Mahoney, 2000: 509). Actors may recognize the analogy between a current problem and previous ones but emphasize a difference that suggests applying a different problem schema. This decision reaches beyond rejecting an invoked analogy as unsuitable, because it points to a particular difference of the case from previous ones and suggests applying a particular alternative problem schema (Schauer, 1987: 594–595). Although the decision is not based on an analogy with previous cases, it is causally connected to them. Actors take previous cases as a reference and decide differently because they are emphasizing a crucial difference. If previous cases resulted in a permissive decision, a reactive sequence will result in a prescriptive one; if previous decisions were prescriptive, it will result in a proscriptive or permissive decision. This chain of temporally ordered and causally connected events is the opposite of a self-reinforcing sequence and creates endogeneity if the initial event has been contingent (Mahoney, 2000: 509). Coordination solutions fulfill this condition. The decision is structured by previous cases even though it does not follow their problem schema. It is distinct from a decision entirely unrelated to previous cases.

Whenever actors trigger a reactive sequence, they create a new class of situations that differ in one crucial dimension from otherwise analogous cases. A decision based on a reactive sequence may provide an analogy for future cases of its type (Schauer, 1987: 572–575) and trigger the self-reinforcing evolution of a new analogy-based rule. Whenever such cases arise, actors can employ the new problem schema. This implies that emphasizing difference halts the expansion of an existing rule in one crucial aspect and creates an additional and initially narrow but potentially expanding rule to cope with a case previously unaccounted for.

Every additional problem schema complements the existing set of informal rules and produces a gradually layered organization. A new case can be addressed without creating structural inconsistency by complementing rather than replacing existing problem schemas. The additional problem schema increases organizational complexity, a frequently observed phenomenon of IOs (Koch, 2009). Accordingly, besides the gradual expansion of existing rules, IOs develop through additional analogy-based rules triggered by reactive sequences. Such institutional layering is well-known in historical institutionalism: “New rules are attached to existing ones, thereby changing the ways in which the original rules structure behavior” (Mahoney and Thelen, 2010: 16). However, the layering process is driven by the stream of non-identical problems that require nuanced responses, not by members struggling to defend an outdated institutional arrangement.

Implications for IOs and the empirical study of analogy-based collective decision-making

Whenever actors adopt an analogy-based collective decision, they recognize the structural similarity between a decision problem and previous ones, irrespective of whether they emphasize similarity or difference. Psychologists (Gentner and Smith, 2012: 132–133; Holyoak and Thagard, 1996) and foreign policy analysts (Houghton, 2001: 26) agree that analogies tend to be suitable if the source and target domains are closely similar in structural terms—that is, deeply connected regarding large relational patterns and not merely regarding similarities with isolated parts or a subset of the matching pattern. For a red car, a white car is likely a better analog than a red football, because carness allows more deep-rooted inferences than redness (Schauer, 1987: 576–579).

Whether actors choose similarity or difference in analogy-based collective decision-making is likely to be affected by the expected effect of an analogy-based decision on the cooperation project. Psychologists and scholars of FPA argue that decision-makers’ current goals are important in analogical reasoning (Gentner and Smith, 2012: 133; Hemmer, 2000: 19–20; Holyoak and Thagard, 1996: 37). The purpose of IO decision-making is to promote mutually beneficial cooperation. Hence, actors will tend to emphasize similarity and follow the existing problem schema of analogous cases if this schema suggests a decision that is broadly compatible with the cooperation project. In contrast, member states are likely to emphasize difference and trigger a reactive sequence if they collectively recognize broad structural similarities between the current case and previous ones, while the existing analogy-based problem schema suggests a decision that threatens to undermine the goal of cooperation.

The theoretical framework suggests that analogy-based collective decision-making produces both organizational resilience and incremental change in the set of rules and routines. An IO will gain resilience if its service to its member states expands and it can increasingly facilitate collective decision-making. When actors emphasize similarity and base decisions on analogy, analogy-based rules expand and the IO can offer solutions for a growing number of decision problems. The expansion will end if other rules, whether based on analogies or not, provide more useful focal points to overcome social ambiguity. When actors emphasize difference and trigger reactive sequences, they support the emergence of new analogy-based rules that offer solutions for additional sets of problems through ever more fine-grained distinctions. Either way, areas for which the IO did not offer solutions are gradually transferred to areas for which the IO does offer routine solutions. Moreover, problem schemas and informal rules for different types of cases are mutually reinforcing. An organizational response is less likely to be subject to pressure toward radical change if it applies to a limited domain, while issues that differ in a crucial dimension are subject to different responses. Members can maintain consistency by creating typologies of similarity and difference (March and Olsen, 1989: 26). The more solutions for different types of situations an IO can offer, the more beneficial is collective decision-making according to its rules, and the fewer are the incentives to resort to package deals or power-based bargaining. This is not the result of an occasional adaptation of formal rules, but of incremental change in the informal rule set through analogy-based day-to-day decision-making.

Nonetheless, there are some limits to analogy-based collective decision-making in IOs. In line with the FPA literature (Breuning, 2003; Houghton, 2001: 28), we suggest that it tends to be used if other mechanisms to achieve collective decisions are unavailable or too cumbersome. Analogies facilitate collective decision-making, but they also constrain actors. Thus, powerful actors might seek to impose solutions on others, or groups of actors might strike package deals, distributing benefits according to the constellation of power. Actors might also base collective decisions on processes of abstract reasoning around the merits of a case, such as evaluating the effectiveness, legitimacy, or fairness of options without referring to previous decisions. Moreover, analogy-based collective decision-making presupposes that actors face social ambiguity: They need to have a general interest in agreement, while differing as to the precise solution. Social ambiguity will diminish if actors gain interest in blocking a decision because it falls outside their win-set, because they have changed their preferences, or because the chosen path has led to policy failure. Note also that powerful members or coalitions may trigger a reactive sequence to halt the undesired expansion of an existing analogy-based informal rule and thus influence the development of the analogy-based rule-set according to their interests. Finally, analogy-based collective decision-making presupposes similar decision problems. It will most likely occur if an IO processes streams of similar cases. Analogy-based collective decision-making most likely occurs in relation to tasks embedded in mutually beneficial cooperation projects, because the general interest may supersede the stakes involved in a single decision. Such tasks include processing exemption requests from Security Council sanctions regimes, classifying endangered species of flora and fauna protected under the CITES convention, authorizing pharmaceuticals, dangerous chemicals or food additives in the EU single market, or financing projects from collective funds in the World Bank. In contrast, isolated, grand or exceptional decisions do not trigger our mechanisms.

Three expectations guide our empirical investigation. First, analogy-based collective decision-making matters for day-to-day decision-making in IOs. Analogies to previous cases and analogy-based informal rules constrain case-specific action even of powerful members. Thus, it renders opportunistic behavior of actors less likely and produces more consistent decisions. Second, analogy-based problem schemas tend to be applied to an increasing range of situations, thus creating drift of informal rules. Third, collective agreement on a reactive sequence leads to layering of informal rules that offer solutions for a growing number of different classes of cases. While the power distribution among members is less relevant when collective decisions are based on analogies, it may well influence the first decision, which defines the problem schema used in analogous cases in the future.

Analogy-based collective decision-making in the Yugoslavia sanctions regime

In this section, we examine the use of analogy-based collective decision-making in the Yugoslavia sanctions committee. The Security Council established the committee to manage a comprehensive trade embargo imposed on Serbia and Montenegro in response to the War in the Balkans (resolutions 713 (1991), 724 (1991) and 757 (1992)). The committee dealt with exemption requests submitted by states, UN agencies, or nongovernmental organizations (NGOs). While each decision was marginally relevant, taken together they greatly impacted the shape and effect of the sanctions regime.

The committee processed exemption requests with broad discretion. It made decisions under a written no-objection procedure and discussed controversial issues during closed meetings (Scharf and Dorosin, 1993: 774, 812). Approval of requests required consensus, which granted all 15 member states veto power, and created the need for horizontal coordination with each incoming case. Nonetheless, committee members were neither obliged to decide consistently, nor did they unanimously favor analogical reasoning as their style of decision-making. Instead, especially the three Western permanent members (P3) demanded that cases had to be decided without consideration of earlier ones, that formal and informal rules did not constrain their discretion, and that every member had a sovereign right to raise objections without internal or public justification (Scharf and Dorosin, 1993: 784). The UK insisted that the committee “operated on a case-by-case basis [. . .],” a phrase used to dismiss earlier decisions as non-binding, because “delegations were constantly reviewing their individual positions” (SR.118, similarly France). The US emphasized that “the Committee had established its rules in such a way as to allow it to change its position over time, should the need arise” (SR.118).

Committee members were in a coordination situation characterized by social ambiguity. The stream of requests constantly confronted them with the problem of agreeing on one solution among several plausible options. On the one hand, committee members agreed on the necessity of the Yugoslavia sanctions regime and wanted to avoid decision deadlock. On the other hand, they had different political objectives (Conlon, 1996: 273; Hannay, 2008: 92). The first group, which included the P3, some other Western members (Austria, Belgium), and Muslim-majority states (Morocco, later Pakistan, Djibouti, and Indonesia), favored strict economic sanctions. The second group, consisting of Zimbabwe, India, Russia, and China, was skeptical toward imposing sanctions (Berdal, 2016: 581). These states criticized the humanitarian consequences (S/PV.3082) and favored broad exemptions. Members of the third group (Japan, Cape Verde, New Zealand, and Chairing delegations of Ecuador in 1992, Brazil in 1993/1994, and Argentina in 1995) were indifferent. The decision process favored the proponents of strict implementation because they had veto power and could block any exemption request. However, proponents were interested in preserving UN members’ support for the sanctions regime. Hence, skeptical members could undermine the sanctions regime through provoking committee stalemate (Conlon, 1995: 652).

Despite the desire of key members to avoid constraints from formal rules and precedents, analogy-based collective decision-making was used to overcome social ambiguity in the committee. Over the course of the 60 meetings evaluated here, members explicitly invoked previous cases more than 110 times, that is, almost twice per meeting. While this seems rare, decision-makers may tacitly follow analogies and tend to explicitly refer to them mainly in non-routine situations. Once a problem schema is established, there is no need to discuss analogies anymore (Houghton, 2001: 28). A broad range of committee members invoked previous cases, including permanent members, mid-size countries, and small states. Sanctions proponents, including the P3, Austria, and Belgium, accounted for almost half of these invocations. Sanctions opponents, mainly Russia and India, also invoked previous decisions to support their positions, albeit less often. Frequently, however, the Chair invoked previous cases to achieve agreement and avoid stalemate. As the Committee met in closed session, analogies were invoked to influence the decision process, not to justify decisions.

To examine whether the mechanisms developed above operate as expected, we analyze a set of unique committee documents. Similar to FPA (Hemmer, 2000: 27; Houghton, 2001), we use process tracing to identify sequences of decisions based on analogy and reactive sequences. Observable implications are (1) the explicit reference by members to previous cases, either invoking previous decisions as an analogy or emphasizing specific difference; (2) decisions made according to recognized previous cases or based on a specific difference from previous cases. Committee documents are taken from the Paul Conlon Sanctions Papers, which comprise original verbatim records of 60 Yugoslavia sanctions committee meetings between 1991 and 1995. ¹

Emphasizing similarity: the expansive logic of informal committee rules on safety exemptions and the transfer of funds to Yugoslavia

Several sequences reveal that the member states overcame social ambiguity by relating otherwise unrelated cases according to structural similarity. Actors emphasized similarity and applied a problem schema to an increasingly broad range of strikingly different situations. This “drift” reflected the expansive logic of analogy-based informal rules.

The handling of safety exemptions is a striking example of relating otherwise unrelated situations according to an abstract-relational criterion, even though the previous decisions were not intended to serve as precedents and powerful member states attempted to resist this development. The sequence started with a permissive decision based on “abstract” reasoning authorizing the Dutch authorities to unload a Yugoslav cargo vessel loaded with coal because “the coal was at risk of spontaneous combustion” (Chair, SR.44; 1992/COMM.1596). Belgium argued that the

basic rule governing the matter was that any ship regarded as originating from the Federal Republic of Yugoslavia [FRY] should be prohibited from loading or unloading cargo, regardless of the origin of that cargo. Very exceptionally, however, and without thereby creating a precedent, requests to unload cargo might be accepted on a case-by-case basis, where a security risk was involved (SR.44, emphasis added), while the vessel and its cargo should be impounded.

The UK and France supported the exemption based on the merits of the case, as “there was a genuine safety problem, [so] they should be permitted to unload the cargo on an exceptional basis” (UK, similarly France, SR.44, emphasis added).

Shortly thereafter, however, the sanctions enforcers struggled to preclude an emerging informal rule on safety exemptions, as sanctions skeptics invoked the analogy to relate a strikingly different situation to the previous case and apply the underlying problem schema. Romania requested to release “empty FRY vessels from the port area of Galatzi on the Danube, where they were creating a dangerous situation due to the freezing of the river” (Scharf and Dorosin, 1993: 805; 1992/COMM.1926; SR.47). The US immediately protested, “since the empty barges [. . .] constituted an economic resource that, under the provisions of the relevant resolutions, could not be returned” (SR.47). Russia, India, and Zimbabwe favored releasing the vessels “to decongest the Romanian ports and for reasons of security” (SR.47). Ultimately, the US accepted the request to overcome social ambiguity “as it had done earlier with respect to the prohibition on providing services to FRY vessels in the [Dutch][. . .] case” (Scharf and Dorosin, 1993: 806), provided that the “present exception does not set a precedent for any future course of action in similar circumstances” (1992/OC.2106, cited in Scharf and Dorosin, 1993: 806, emphasis added).

However, in effect, the committee had already accepted a rule-like “safety exemption” and even powerful sanctions enforcers could no longer limit its drift to other situations (Scharf and Dorosin, 1993: 806). Romania requested to ship “fuel to ensure the functioning of FRY ice-breaking vessels on the Danube River,” which the committee accepted based on the previous safety exemptions (Scharf and Dorosin, 1993: 806). Romania further requested to ship fuel to Yugoslavia “to ensure unimpeded operation of the Iron Gates hydroelectric and navigation systems during the coming winter” (1993/COMM.12587). While the US opposed this, the Chair emphasized similarity and stressed that “the Committee had considered a similar request earlier [. . .], which it had approved on an interim basis” (SR.89). The US finally agreed to the shipment of fuel provided that it was only for one month (SR.90). Yet, the next winter, Romania repeated the request to ship fuel to Yugoslavia for the Iron Gates plant “to ensure its functioning throughout the winter” (1994/COMM.48436; SR.113) and emphasized similarity: “The Committee had authorized such requests in the past” (SR.113). The US again tried to avoid the implications of this analogy and argued that “in her Government’s view [. . .] Yugoslavia (Serbia and Montenegro) had sufficient fuel to meet its basic needs” (SR.113). However, Russia, emphasizing similarity, demanded “that the Committee had released fuel for Iron Gates the previous two winters and should do so again, to ensure that the possibility of accidents [. . .] would be avoided” (SR.113). Eventually, the US accepted the request (United Press International, 1995; S/1996/946, para. 38).

The handling of fund transfers to Yugoslavia also illustrates how analogy-based informal rules are self-reinforcing and drift over time, while they do not always evolve in a linear fashion. The Security Council had sanctioned the transferring of funds or economic resources to Yugoslavia, “except payments exclusively for strictly medical or humanitarian purposes and foodstuffs” (resolution 757 (1992), para. 5), however, without defining the exact meaning of this provision (Scharf and Dorosin, 1993: 787–788).

The first decision was power-based and proscriptive, albeit contested. Australia (1992/COMM.16) had requested a monthly payment of $750,000 in “Australian pensions to Australian citizens residing in Serbia and Montenegro.” The US used its power to veto the request. It argued that “the intent of the resolution” was to prohibit payments “not exclusively for strictly medical or humanitarian purposes” and cautioned not to “create a significant loophole” (SR.9). In contrast, all other members favored accepting the request (SR.9). The Chair urged that “the Committee should not prejudge what action it might take on similar cases in the future” (SR.9), and Belgium hoped that “future cases would be [decided] on the basis of the specific circumstances of each case, the more specific such requests were, the easier it would be to approve them” (SR.10), thus suggesting a possible reactive sequence based on the difference between general and specific requests.

More specific requests triggered a permissive and expanding analogy-based informal rule. Greece requested (1992/COMM.18, Add.1) to exempt “Greek students pursuing degrees in Yugoslav universities” (Chair, SR.9). The US, UK, France, and Belgium initially resisted (“too general”), but the committee finally approved the request if remittances were “confined to students who were about to complete their studies” (Chair, SR.11). Subsequently, the committee applied the permissive problem schema to a Swedish request (1992/COMM.227) to exempt an alimony payment (US, SR.23), which “corresponded to the position already taken [. . .] in similar cases” (Chair, SR.23), and a Dutch request to exempt monthly remittances for 150 orphaned Yugoslav children (1992/COMM.307), although the US insisted that “its decision did not prejudge its position on other matters of humanitarian concern” (SR.27). When a Yugoslav NGO requested to exempt “all social security payments” made under bilateral agreements (1992/COMM.1022/1092), the US, in view of 14 members approving and similar cases accepted before, retreated to the position that states could but were not obliged to pay pensions to persons in Yugoslavia (SR.41; S/1996/946, para. 64). According to the logic of expansive and drifting informal rules, the US henceforth generally acquiesced to pension remittances although they had vehemently opposed such payments.

Reactive sequences: creating novel solutions for the repatriation of goods and for diplomatic goods

Several episodes demonstrate that analogy-based collective decision-making may result in reactive sequences and the desire to adopt a decision to the contrary. These decisions are related to previous ones but do not follow the same problem schema. Instead, they offer novel solutions for a new class of cases. As expected, the new solutions are generally in line with the spirit of the underlying cooperation project. Without creating inconsistency, they are layered within the set of existing solutions.

Regarding the repatriation of goods remaining in Yugoslavia but owned by third countries (Scharf and Dorosin, 1993: 796–798), the committee first developed a permissive informal rule. The first case concerned a “flight of a United Kingdom aircraft from Belgrade” (Chair, SR.14, 1992/COMM.68), which the committee granted based on abstract reasoning. Shortly thereafter, it dealt with a French request (1992/COMM.166) to grant the “return from Belgrade to Paris of an aircraft owned by a French enterprise” (SR.17). The Chair invoked the analogy and the committee granted the request accordingly because it “had previously approved a similar request” (SR.17). Later, the committee approved a request to repatriate a Turkish aircraft, “which had been sent to Yugoslavia for maintenance” (1992/COMM.197; Chair, SR.21) prior to resolution 757 (1992) by emphasizing similarity “since it would be a simple case of repatriating an asset belonging to that country” (UK, SR.21).

However, driven by the US intention to limit the expanding informal rule, the committee introduced a crucial distinction. Mongolia (1992/COMM.105) had requested to export “a quantity of hotel furniture and related equipment purchased by the Government of Mongolia,” because the contract “had been concluded and the money transferred [. . .] prior to the adoption of resolution 757 (1992)” (Chair, SR.12). Japan, India, Russia, China, Austria, and Hungary advocated approving the request because the resolution “prohibited exports ‘after the date of the present resolution’” only (Hungary, SR.18, emphasis added). However, the US insisted that an approval “would be setting a precedent that might result in the submission of many similar requests” (SR.27). In effect, the US used its veto power to limit the repatriation of goods in future cases (SR.28 and Chair’s reference to decision, see SR.74). This move established a second but connected problem schema and created a layered set of complementary informal rules guiding two sets of remarkably consistent committee decisions. The US invoked the Mongolian analogy to reject a Norwegian request to export “a ship commissioned by Norway” from Yugoslavia, as the decision made “with respect to Mongolia's request also applied to Norway's request” (SR.27). The US and UK also rejected a French request (1992/COMM.438) “for fruit to be exported from [. . .] Yugoslavia for delivery to a French company [. . .], which had exported merchandise to Serbia on account against the following summer’s fruit crop” (Chair, SR.30) “for the same reasons that it had previously opposed the export of furniture to Mongolia” (US, SR.30) and because the request “clearly involved the export of goods from Serbia” (UK, SR.30). The committee also rejected an Egyptian request (1992/COMM.582/Rev.1) “to repatriate to Egypt equipment apparently sent to [. . .] Yugoslavia for repairs” (Chair, SR.31) because the repairs “had added value to the equipment, thus making the matter more than a simple case of repatriation” (UK, US, Austria, SR.31). The committee further dismissed a Bulgarian request “for delivery [. . .] of two river barges that had been contracted for before the imposition of sanctions and which had already been paid for” (1994/COMM.30131) because “in a similar matter, the Committee had declined to approve the export [. . .] of a vessel built for a Norwegian firm” (Chair, SR.105).

Conversely, the committee approved a request from Uruguay to return raw fabric that “had been manufactured in Uruguay for a United States company and had not been used or altered in Yugoslavia,” which it had done “in similar cases, as long as no transformation of the goods had occurred and no payment had been made” (Chair, US, SR.33). The committee also granted two UK requests for repatriating “medical products belonging to a United Kingdom company which were currently stored in a warehouse at Belgrade” (1992/COMM.1130, also “three rail tankers,” 1992/COMM.1137, SR.37). It approved a French request (1993/COMM.4149) to return an electric motor purchased in March 1991, which “in accordance with an after-sales contract [. . .] had subsequently been returned to the manufacturer [. . .] and no financial transfer was involved” (Chair, SR.74). The US acquiesced because

if no financial transaction was involved [. . .] the issue was solely one of the repatriation of French property, [. . .] provided that the Committee was not acting in a manner inconsistent with any precedent it might have set in the Mongolian case. (SR.74)

Ultimately, the committee tacitly formalized the two mutually complementary and layered informal rules. The Chair recalled the two informal rules:

While the Committee had consistently approved the repatriation [. . .] of property belonging to other States when such property was not of Yugoslav origin, he recalled a case involving hotel equipment purchased from Yugoslavia by Mongolia before the sanctions had come into effect, the transfer of which to Mongolia the Committee had been unable to approve. (Chair, SR.74)

In its annual report, the committee noted that it had “formulated general principles regarding the repatriation of goods” (S/25027, para. 17), which “focused on a specific legal aspect” (Committee Secretary, SR.74), so that such goods could be exported if they did not originate in Yugoslavia, if their ownership had never been transferred to Yugoslavia, if they had not been altered in Yugoslavia, and if their use was never under Yugoslavian control (S/25027, para. 17).

The committee’s handling of diplomatic goods exemplifies once again how emphasizing a crucial difference that triggers a reactive sequence may avoid undesired decisions undermining the goal of cooperation caused by drifting problem schemas. Committee decisions generally followed a permissive informal rule. When Greece inquired “as to whether the embargo applied to imports and exports for the sole need of embassies, diplomatic and consular missions” (Chair, SR.10), committee members accepted the need for diplomatic interaction, even if such exemptions might provide avenues to circumvent the embargo (Belgium, SR.10). Applying this problem schema, the committee approved two Yugoslav requests for shipments of the personal effects of its diplomats abroad (Chair, SR.13, SR.14) and further expanded the rule to funds for Yugoslavian embassies (1992/COMM.83, SR.14/16). Later, it extended the applicability to the repatriation of paintings and sculptures from the Yugoslav Press and Cultural Centre in New York (1992/COMM.395), which the committee regarded as an “official agency” (Chair, SR.31).

However, when Yugoslavia requested to “repatriate art works which had been on display at Yugoslavia's Cultural and Press Centre in Paris” (1992/COMM.1276), the committee introduced a crucial distinction to avoid undesired drift. Croatia (not a member of the committee) requested that “the application to ‘repatriate’ those materials to Yugoslavia should be denied” because it claimed ownership of the art works, which had fallen into Yugoslav hands during the occupation of the Croatian city of Vukovar. When India insisted that on “a previous occasion, the Committee had authorized the return of certain cultural property to Yugoslavia,” France objected (SR.41). The Chair emphasized the crucial distinction, namely that “in that case the ownership of the property had not been contested” (Chair, SR.41; also S/25027). Hence, the Committee triggered a reactive sequence. It rejected the request based on a new proscriptive problem schema for this new type of case.

Definition of conditions to limit the expansive logic of informal rules: Yugoslav athletes at sporting events

The handling of Yugoslav athletes at sporting events (Scharf and Dorosin, 1993: 810–811) exemplifies how decision-makers may seek to limit the undesired drift of self-reinforcing informal rules by committing themselves to precise conditions of applicability.

When the International Olympic Committee (IOC) requested “the participation of a team from [. . .] Yugoslavia” in the Barcelona Summer Olympics under the IOC flag and in white uniforms (SR.22), an intense dispute arose over the similarity to, or difference from, a previous case. The committee had previously considered a Filipino request and adopted a proscriptive decision preventing a Yugoslav national team from participating at the 1992 Chess Olympiad in the Philippines in June 1992 (SR.10) because the Security Council had required all states “[. . .] to prevent the participation in sporting events on their territory of persons or groups representing [. . .] Yugoslavia” (resolution 757 (1992), para. 8b). In the Barcelona case, the Chair directly invoked the analogy: The committee “had considered a similar request from the Philippines [. . .]. That request had been denied because it was viewed as an attempt to circumvent Security Council resolution 757 (1992)” (SR.22).

Sanctions proponents sought to establish similarity between the two cases. Austria staunchly opposed the IOC request and referred to the analog:

It was clear that the games involved national pride. The fact that the Yugoslav Government had campaigned for its athletes' participation testified that Belgrade held such an interpretation. The case involving the chess tournament was very similar [. . .]. It could be claimed that the difference [. . .] was that IOC, rather than the Yugoslav authorities, would select the team members participating, but that was clearly a fiction. (SR.22)

Hungary also opposed the request (SR.22). The Chair suggested a negative response and demanded a “consistent approach” because the committee “had already decided that participation by Yugoslav players in a chess tournament under a formula similar to” the IOC proposal “was a way to circumvent a Security Council resolution” (SR.22). The UK agreed that the request should be rejected as “a form of official participation in disguise” because “the athletes would travel together, be housed together, and parade together” (SR.23). Ecuador recalled that “one of the Committee's first decisions had been to refuse to authorize the Yugoslav team to participate in the Chess Olympics [. . .]. The Olympic Games were an event in which national teams, made up of athletes representing States, participated” (SR.23).

In contrast, sanctions sceptics emphasized a crucial difference. Cape Verde maintained that there was “a slight difference” between the two cases because “in the Philippine case, authorization had been requested for [. . .] a national Yugoslav team, which was not the case for the Barcelona Olympics” (SR.22). Russia argued that

the analogy drawn by [Ecuador] was incorrect in that, in the case of the Chess Olympics, the team officially represented Yugoslavia, whereas in the case of the Barcelona Olympic Games, the athletes would be selected and sent by IOC and would not represent any Government. (SR.23)

Even France maintained that “it was understood that the athletes selected by IOC would not be representing Yugoslavia” (SR.23). India agreed that “a distinction should be made between the two events” (SR.23).

The Chair summarized that

some delegations had referred to the case of the Chess Olympics [. . .]. The Committee could follow the same course [. . .] but there was a new element: [. . .] it would not represent any form of Yugoslav State, country or nation and the final list of athletes and officials [. . .] would be established by IOC. (SR.23)

Some countries indicated that they accepted the distinction, provided that some conditions were fulfilled. France and the US signaled they would acquiesce to the IOC proposal if athletes competed individually and were not representing Yugoslavia (SR.22).

Eventually, the committee abandoned its strict stance, approved the participation of Yugoslav athletes in the Olympic Games, and intended to limit expansion and drift of the resulting permissive problem schema by “a series of conditions under which athletes might participate in international sporting events”: (1) The athletes should not appear together or participate as part of a team, nor should they appear in any uniform, under any banner or flag, or have any anthem dedicated to them. (2) The sporting organizations had to select participating athletes. (3) The athletes might be accompanied by coaches and assistants, but not by officials. (4) The athletes had to avoid making any political statements (S/25027, para. 18–19). This excluded team events, because “participation in team events would inevitably evoke representation of their country” (Scharf and Dorosin, 1993: 811).

Subsequently, the Barcelona Olympics decision served as an analogy for further cases, and the committee applied the conditions with the utmost strictness. It granted a Finnish request to allow skaters from Yugoslavia to participate “as individuals” in the 1993 Figure Skating Championships (1992/COMM.1657) after Austria and Hungary referred to “a similar request [. . .] regarding [. . .] the Barcelona Olympic Games” and insisted on “the four conditions.” Austria “took the view that any event involving the participation of more than one athlete, including events such as pairs figure skating, constituted a team sport and should thus not be permitted” (SR.44). In line with the analogy, the committee allowed the participation of individual athletes only (SR.44). The handling of participation in sports events even became routine and the solution was consistently applied to the rowing championship in Bulgaria (SR.34, 1992/COMM.725), a table tennis tournament in Las Vegas (SR.46), and the Atlanta Olympic Games (SR.108, 1994/COMM.38805).

Conclusion

Analogy-based collective decision-making in IOs produces an endogenously emerging and self-reinforcing set of informal rules that is both firmly established and changes incrementally. Two complementary micro-level mechanisms drive this path-dependent process. If collective decisions emphasize similarity between a current case and a former one, informal rules emerge, become increasingly established with every additional case, and gradually turn non-routine matters into routine ones. These rules evolve incrementally because cases are never identical. If actors recognize the analogy to previous cases but reject the solution due to a crucial difference, they establish a new problem schema for dealing with cases of this class. Accordingly, the organization can solve an increasingly broad range of situations and gains resilience.

The Yugoslavia sanctions committee demonstrates that analogy-based collective decision-making significantly affected member state choices, helped overcome social ambiguity, and followed the two mechanisms. Despite a prevailing conflict, member states frequently invoked analogy and collectively derived problem schemas from previous cases—after an initial decision based on the merits of the case or on member state power. First, members repeatedly emphasized similarity in strikingly different situations, such as unloading a coal barge at Rotterdam and supplying fuel for a hydroelectric power station at the Romanian border. Firmly established informal rules emerged from streams of cases and drifted to non-identical cases considered to be similar. These rules even constrained powerful members, which could have vetoed any undesired approval. Declaring a decision to be exceptional did not preclude its use as an analog, but specifying the conditions had some effect. Second, members repeatedly triggered reactive sequences to limit the expansionary logic of analogy-based rules. They created distinctions that treated seemingly similar cases differently. For instance, the repatriation of foreign property located in Yugoslavia was generally accepted. However, it was rejected if activities in Yugoslavia had added value to it. Consequently, the committee developed a dense web of layered informal rules that provided an increasingly nuanced set of routine solutions to different problems.

Existing approaches have difficulty explaining these findings. Although our argument assumes strategic actors, standard rational choice approaches and intergovernmentalism cannot explain why self-reinforcing informal rules emerge endogenously through analogical reasoning and affect collective behavior even of powerful members. The conventions concept accounts for emerging informal rules but cannot grasp incremental change. Sociological institutionalism would expect rule-based behavior to be rooted in rule internalization. However, it has difficulty explaining why informal rules emerge among strategic actors and why actors occasionally reject seemingly rule-conforming decisions.

Being firmly grounded in a rationalist strategic action framework, our concept demonstrates the merits of combining historical institutionalism with analogical reasoning approaches to understand IO resilience and incremental change through day-to-day operations. Historical institutionalism emphasizes the sequencing of decisions as an endogenous factor for institutional development over time, resulting in drift or layering processes. It also stresses the importance of elucidating the micro-level mechanisms that produce self-reinforcing rules and reactive sequences. Psychological concepts of analogical reasoning, on which FPA analogical reasoning approaches draw, direct our attention to how actors choose suitable analogs and develop schemas for problem-solving through analogies. The combination of these concepts and their application to the analysis of day-to-day IO decision-making reveals the importance of analogy-based collective decision-making for IO resilience and incremental change.