Hard International Law-Contributing Organizations as Networks

1 Introduction

Which intergovernmental organizations (IGOs) generate new international norms and laws and regulate and enforce state compliance with treaties and other obligations? This question is important because it relates to IGO efficacy in promoting cooperation.^[1] To regulate state behaviors and enforce compliance, IGOs must be equipped with the necessary institutional mandates and mechanisms. Moreover, international law-oriented IGOs and their member states relate to one another to create an international legal structure that may ultimately affect conflict and cooperation in the international system. Both legal scholars and political scientists study international law and international organizations, but a limited amount of research joins these two topics (Alvarez 2005; Slaughter-Burley 1993). There has been little systematic study about how the institutional design of IGOs relates to their ability to be agents and creators of international law using a large sample of IGOs. The goal of this project is to map systematically IGOs’ involvement with “hard” international law, which requires both a higher level of state commitment to multilateralism and the independence of IGOs. This project presents new data on IGOs’ institutional characteristics to help us understand which organizations are designed to generate and implement international law. In this paper we explore both how these “highly legalized” IGOs relate to each other and how states, through their membership in such IGOs, are connected to each other in the creation of a complex international law network.^[2] These data should be important for explaining many patterns of cooperation and conflict between states, as well as compliance towards international agreements.

Because the structure of IGOs, including those features related to international law, and their ability to influence state behavior are endogenous to their member states’ willingness to be affected by such institutions, we employ a network-analytic approach in this investigation.^[3] Scholars interested in compliance with international law are increasingly adopting network arguments, suggesting that states’ networks of shared IGO membership promote compliance with various international legal provisions (Goodliffe et al. 2012; Greenhill 2010; Prorok and Appel 2014; Von Stein 2008, 2010). Thus, we approach IGOs’ role in the creation and effectiveness of international law in terms of two interrelated networks: (1) a broader network that connects IGOs to one another as a function of their involvement with international law; and (2) an affiliation network between states resulting from their shared membership in international law-involved IGOs.

While the data collection project is ongoing, this paper offers a first look at the data for the 1980–2005 period. The data are “biased” towards currently existing IGOs, but they allow us to provide a first inferential investigation of the institutional features and functions that drive IGO contribution of hard international law, as well as a descriptive exploration of the resulting networks of what we label as highly legalized IGOs and state membership in such networks. This effort helps us raise new hypotheses about the influence of IGOs on peace and conflict although, ultimately, the data collected in this project could be used to study the effects of IGO networks on many forms of state behavior.

2 Linking international law and IGOs in the literature

The systematic study of international institutions has grown considerably over time. Two main approaches have emerged, one focusing on individual IGOs and the other employing samples or the whole population of IGOs. Here we take the latter approach but borrow theoretically from the rational institutional design literature that is often employed to study single IGOs. The literature that examines IGO attributes and their effects on state behavior is too extensive to be adequately reviewed here, so we focus more narrowly on studies pertaining to institutional design and/or international law especially those that involve samples or the population of IGOs.^[4]

Several studies and large N data projects have provided systematic classifications of IGO institutional attributes and mandates (Boehmer, Gartzke, and Nordstrom 2004; Cox and Jacobson 1973; Haftel 2007; Haftel and Thompson 2006; Hooghe and Marks forthcoming; Jacobson 1984). The IGO institutional score and mandate data produced by Boehmer, Gartzke, and Nordstrom (2004) –a partial source of data for this project– has been used in others studies (Bearce and Bondanella 2007; Boehmer and Nordstrom 2008; Donno, Metzger, and Russett 2015; Dorussen and Ward 2008; Hansen, Mitchell and Nemeth 2008; Ingram, Robinson and Busch 2005). Few of these works, however, explicitly connect IGOs’ institutional features to international law. Similarly, a number of studies directly looks at IGOs in their attempt to evaluate the effectiveness of international law by investigating states’ compliance with specific legal provisions, often concerning conflict behavior or human rights (Goodliffe and Hawkins 2006; Goodliffe et al. 2012; Greenhill 2010; Huth, Croco, and Appel 2011; Morrow 2007; Prorok and Appel 2014; Simmons 2002; Valentino, Huth, and Croco 2006). Some of these studies argue that the effectiveness of international law should be assessed based on states’ shared membership in IGOs, as IGOs may either directly enforce compliance with international law or provide third party states with additional tools to promote compliance (Goodliffe et al. 2012; Prorok and Appel 2014; Von Stein 2010). Our goal in this project is to connect IGO attributes and their embeddedness to the legalization conception of international law, based on the foundation provided by Abbott and Snidal (1998), Abbott et al. (2000), Koremenos, Lipson, and Snidal (2001), and Hawkins et al. (2006). The next section provides a brief discussion of our theoretical foundation for this data project.

3 Theoretical propositions linking IGO attributes to international law

We begin from the premise that IGOs are the products of states’ preferences (Koremenos, Lipson, and Snidal 2001) and are rationally formed to produce collective goods and reduce transaction costs, but they also help to create and enforce new rules and norms (Abbott and Snidal 1998, 8). For states to accomplish the goals for which they become active in IGOs, these institutions require some degree of centralization and independence (Abbott and Snidal 1998).

States’ decisions about which institutions should be given greater functional capacity is not random. We solve this issue by linking IGOs to the degree of legalization provided by states. States use IGOs and international law, through international treaties, to legalize norms and patterns of behavior that they see in their interest (Abbott et al. 2000). The extent of legalization of IGOs relates to their ability to contribute to hard or soft law. The degree to which individual IGOs are connected to hard international law (O, P, D) depends on: differences in how states are obligated (O) to comply with international law; the precision of the law (P); and the delegation (D) of authority to third-parties, in this case IGOs. States ratify hard law treaties –which are often formed through IGOs– and create IGOs to help enforce international law. The creation of precise, binding international laws, and the delegation of authority to IGOs are meant to limit state sovereignty and facilitate the identification of non-compliant behavior. Members of hard-law based IGOs signal a public commitment to certain norms and rules in the interstate system.

The quantity and quality of state IGO memberships matter and, although states may not have set out to design a broader international law network, one can be said to emerge as states join the same IGOs to advance hard international law. The relationship between states’ embeddedness and cooperation is endogenous. Shared ties in a network of IGOs reflect states’ preferences for seeking cooperation by creating and enforcing rules in the international system, some akin to covenants that should not be broken for short-term interests. “Most international agreements are simultaneously contract and covenant.” (Abbott and Snidal 2000, 425)

The embeddedness of states within the network of hard-law connected IGOs reflects a commitment to solve interstate conflicts through third party institutions. These states should be more likely to adopt behaviors that are not just contract-based, but internalize deeper norms and values. We expect states that are embedded in the network of hard-law IGOs to enjoy a higher level of cooperation.

To refine further the prediction that hard law related IGOs contribute to cooperation, we classify IGOs based on Young’s (1999) typology. Young categorizes international regimes by mission types: procedural, programmatic, regulatory, and generative. The latter two mission types relate to international law and IGOs. IGOs with regulatory missions help regulate and monitor behavior–e.g. the IAEA and the WTO. We hypothesize that state dyads most embedded in the networks of IGO with regulatory mission are more committed to the enforcement of international law and enjoy interstate peace compared to less embedded state dyads. Similarly, we expect the same type of shared preference for international law for state dyads most deeply tied to highly legalized IGOs with a generative mission as these IGOs seek to expand and deepen international law.

4 Methods

The data for this project come from the IGO attribute dataset (Boehmer 2011), which provides many measures of IGO institutional characteristics, including a new measure of IGOs contribution to hard international law, labeled “IGO Hardness.” and new variables reflecting Young’s (1999) regime typology. We examine relationships between IGO attributes and their contribution to international law as a function of their mission and map out the network of highly legalized IGOs for the 1980–2005 period. The population of IGOs used for this study is based on the revised Correlates of War Intergovernmental Organizations (2.3) dataset (Pevehouse, Nordstrom, and Warnke 2004), which includes 12,856 IGO years. Each IGO is measured for each year of its existence.

The IGO Hardness variable is a composite index that ranges from zero to five, based on an IGO’s possession of five key institutional attributes: Formal Voting, Operations Codified, Bureaucracy, Binding Decisions, and Juridical mechanism. The higher the value of this variable, the more it should capture the extent to which an IGO is related to hard international law, whereas low values relate the IGO to soft international law based roughly on Abbott and Snidal (2000) continuum. We label IGOs that score high –four or more– on this variable as “highly legalized.” We also create four binary variables that classify IGOs’ mission based on Young’s typology: regulatory, programmatic, procedural, and generative.^[5] IGOs are also coded as possessing specific issue area focus: economics, security, or matters of culture/norms.

We first present descriptive statistics about the IGO traits defined in the above-mentioned Young typology. We then focus on the IGO Hardness index and use network analysis to look at IGOs from a relational perspective. We look at the IGO data first as one-mode networks in which IGOs constitute the nodes and are related to one another by sharing a Hardness scale of four and above. This approach provides a picture of the “thickness” –the density– of the network of those IGOs that are expected to mostly influence international law and state behavior. We then shift our focus to member states and look at the data as two-mode (affiliation) networks in which states are connected by shared membership in IGOs with a high degree of hardness.

5 Exploratory empirical findings

Figure 1 presents the frequency distribution of international organizations with regulatory, programmatic, procedural, and generative function over time. Most IGOs in any given year have programmatic and/or procedural function. A minority of IGOs have regulatory or generative functions, which we expect to contribute more strongly to the development and enforcement of international law. Note also the overall limited longitudinal variation in the data. Few IGOs add or shed functions during this limited period of time, and whatever variability is observed is due to the “death” of organizations.^[6]

Figure 1:

Distribution of IGO function according to Young’s (1999) typology.

We investigate which institutional features better predict specific IGO missions according to Young’s typology. We ran a series of logit regressions in which the dependent variable is whether or not an IGO takes on a regulatory, procedural, programmatic or generative mission. The results are presented in Table 1 below. All of the five institutional attributes that we have identified in the previous section as key determinants of high legalization – formal voting, operations codified, bureaucracy, binding decisions, and juridical mechanism– are positive and significant for regulatory and generative IGOs, whereas the procedural and programmatic IGOs show some attribute coefficients that are negative and/or insignificant.

The other IGO trait we are mostly interested is IGO Hardness. The dot-plot in Figure 2 below shows the frequency distribution of IGOs based on their hardness index during the 1980–2005 period. The diameter of each dot is proportional to the number of IGOs with a certain hardness score. Figure 2 shows that most IGOs are relatively bare bone operations with few hard operational features. Only a minority of IGOs achieve a score of 4 or higher and can be considered highly legalized and the traits least frequently present are the ability to make binding decisions and juridical mechanisms. This may indicate that IGO contributions to hard international law comes from a rather limited number of organizations. States make stronger commitments to international law by joining these IGOs.

Figure 2:

Frequency distribution of the hardness index across IGOs.

We use the Hardness variable, in addition to whether IGOs have an economic or security focus, in a series of logit models to explore whether they assume regulatory, programmatic, generative, or procedural mission. The results are presented in Table 2. IGOs with a security mandate are positively and statistically related to the presence of a generative or regulatory mission. Serious security commitments between states appear to positively relate to state ties to hard international laws through regulatory and generative IGOs. Having an economic focus has a more ambiguous effect, actually reducing the likelihood of assuming a law-generating function. Hardness emerges as a consistent predictor of regulatory and generative missions.

We employ descriptive network analytical techniques to map the network of law and behavior-influencing IGOs. Because the structure of the emerging networks varies little over time, we provide only the snapshot the year 2005. In this preliminary analysis individual IGOs represent the nodes in the network, and we take the edges –the links between them– to be the possession of a hardness index of 4 or above. Figure 3 shows the graphical representation of influential IGOs in 2005. In the graph each IGO is labeled according to its Correlates of War name. Not surprisingly, this network is constituted by some of the most established and popular global IGOs –such as the UN, the IMF, the WTO, ILO, and the ITU– and by some of the most widely known regional IGOs –such as the EU, the OAS, CARICOM, ASEAN, and ECOWAS.^[7]

Figure 3:

High hardness IGO network in 2005.

The goal of mapping the network of influential IGOs is to derive broader structural attributes of the networks themselves that may play a role in shaping state behavior and the creation of international law. For each data-year we derive some standard descriptive statistics for these networks which may be used as predictors of the extent to which states are affected by the web of IGOs and law in which they find themselves immersed. Table 3 below shows the number of nodes and edges, as well as density and centralization of the highly legalized IGO network in each year. Changes in the number of nodes reflect changes in the general number of IGOs in the system, whereas changes in the number of edges reflect variation in the number of IGOs that become highly legalized. Density reflects the “thickness” of the highly legalized IGO network, while centralization expresses the extent to which the network revolves around specific set of highly central IGOs.

We also look at highly legalized IGOs as affiliation –two-mode– networks in which states are the nodes and are connected by shared membership in high hardness IGOs. We created a weighted affiliation matrix the cells of which contain the number of high hardness IGOs shared by state dyads. We do not provide a graphical representation of these networks because the resulting graphs are too dense to display any clear pattern of association. However, we can still extract very useful statistics at the network and dyad level that further illustrate how scholars may benefit from these data to explore a variety of propositions about the role of IGOs in shaping state behavior and in creating international law.

In Figure 4 below, for example, we compare the distribution of dyadically shared ties in all IGOs, regardless of the type, with the distribution of shared ties in highly legalized IGOs only. We note that, of course, the average number of shared ties in highly legalized IGOs is considerably smaller than the average number of shared ties for all IGOs, but that the distribution of shared ties in highly legalized IGOs is considerably more normal and less skewed than the distribution resulting from all shared IGO memberships. This should have observable consequences on the estimation of models that explore the impact of shared IGO membership on state behavior.

Figure 4:

Frequency distributions of dyadic ties for all IGOs and highly legalized IGOs.

6 Discussion and conclusions

This paper provides a starting point to begin thinking about how IGO attributes relate to international law. Our results show that our Hardness measure predicts IGOs with regulatory and generative missions. The results are encouraging and suggest that IGOs’ institutional traits are important factors in capturing differences between individual organizations both in terms of their ability to contribute to the creation of international law and in terms of their mission. In summary, we speculate that states highly tied to the networks of highly legalized IGOs with regulatory or generative mission share a high degree of commitment to international law, and should be less likely to militarize their conflicts. The next steps in the project are to complete the data collection so that the hypotheses raised here can be tested and, in the longer run, to explain the state attributes that lead them to become embedded in this international relations network and to use these network ties and connections to explain various forms of state cooperation and compliance.