  |
[*pg 151]
WHEN, WHAT, AND WHY DO STATES CHOOSE TO DELEGATE?
I. INTRODUCTION
II. RATIONAL DESIGN THEORY
III. DATA
IV. DESCRIPTIVE STATISTICS
V. FINDINGS: WHEN IS DELEGATION MOST LIKELY?
A. Delegation and the Complexity of the Cooperation Problem
B. Delegation and Number and Heterogeneity of Parties
C. Other Possible Explanatory Variables
D. Relationship Between Delegation and Other Design Provisions
E. Multivariate Analysis
VI. CONCLUSION
FOOTNOTES
 |
I. INTRODUCTION
The introductory article in this symposium offers a rich conceptual framework for understanding international delegation. Curtis A. Bradley and Judith Kelley make great strides in variable conceptualization, defining and elaborating some key dimensions of delegation and thereby serving as a very useful springboard for more refined theoretical and empirical work.1 Many of the other articles in this symposium provide case-study evidence of delegation in some major agreements, with some of them richly examining the consequences of delegation for the implementation of the agreements.2
Employing a random sample of international agreements, this article is the first systematic look at when, what, and why states choose to delegate.3 As the descriptive statistics below indicate, delegation is widespread, and, as the theory elaborated here would predict, it is more likely to be incorporated into [*pg 152] agreements when states are trying to solve complex cooperation problems and when larger numbers of states are involved in the cooperative endeavor. Throughout this article, internal delegation is defined as delegation to a collective formed by the members of the agreement themselves, as distinguished from external delegation, defined as delegation to a third party outside of the agreement. This distinction turns out to be quite relevant.
Other interesting, and often surprising, findings emerge. For example, given this initial probe into the data, delegation is evidently not limited by factors like hegemony or risk aversion, whereas democracy is negatively correlated with delegation. Additionally, this article analyzes the correlation between delegation and other dependent variables -- that is, other institutional design variables that are under the control of the states making the agreement. Both withdrawal clauses (also known as exit or denunciation clauses) and finite duration provisions are significantly correlated with delegation. The positive correlation between withdrawal clauses and delegation suggests that states are minimizing their sovereignty costs by creating an outside option. The negative correlation between finite duration provisions and delegation suggests that delegation can function as a form of flexibility, thereby acting as a substitute for finite duration. This article confirms the importance of international delegation as a topic of focused study by documenting it as an important and nontrivial empirical phenomenon.
The empirical work in this article is anchored in the theoretical framework of Rational Design.4 The starting point for Rational Design is a very simple observation: institutionalized international cooperation is organized in significantly diversified ways.5 More important, however, is the theoretical reason for this observation: different international institutions are solving different combinations of cooperation problems. This implies that differences among international institutions are not random; rather, states and other international actors shape institutions to solve the specific problems they face. In other words, design variations are largely the result of rational, purposive interactions. The goal of the Rational Design volume is to offer a systematic account of five design variables (membership, scope, centralization, control, and flexibility -- the dependent variables in the framework), relating them to recurrent cooperation problems faced by states, the independent variables.
This article uses Rational Design to explain the institutional-design choice of international delegation, where delegation is equivalent to the rational-design variable of centralization. A state delegates functions to an international body if the expected benefits from delegation outweigh the expected costs. [*pg 153] Empirically, the extent and type of delegation should vary according to the importance of delegation in solving the particular cooperation problems faced by states. From this perspective, the debate on whether delegation is an important part of international law (not the perspective emphasized in this special issue) is misplaced. A more fine-grained theory predicts that delegation is important when the underlying cooperation problem requires it.
Rational Design theory is sketched out in Part II. The data are presented in Part III, followed by a detailed set of descriptive statistics in Part IV. The theory is then tested in Part V, and Part VI concludes and points to new directions for research once more data are collected.
|
II. RATIONAL DESIGN THEORY
Given its primacy in Rational Design theory, what exactly is a cooperation problem? One way of conceptualizing the notion is to focus on the categories of interests and constraints. Interests are captured in two of the Rational Design independent variables, with the enforcement problem referring "to the strength of individual actors' incentives to cheat"6 and the distribution problem depending "on how each actor compares its preferred alternative to other actors' preferred alternatives."7 Constraints are captured with variables (italicized below) that address both information and beliefs: uncertainty about preferences (that is, uncertainty regarding what the other state partners' preferences are), uncertainty about behavior (not being able to decipher easily whether the other state partners are cooperating or defecting), and uncertainty about the state of the world (that is, uncertainty regarding the consequences of cooperation -- who will benefit the most, et cetera). To these, a few more variables can be added, including the commitment problem (defined as a domestic commitment problem or a time-inconsistency problem).8
With respect to international delegation, the set of rational-design conjectures surrounding the dependent variable centralization (defined as whether institutional tasks are performed by a single focal entity or not) are pertinent.9 Of sixteen suggested univariate10 conjectures relating one [*pg 154] independent variable to one dependent variable, four conjectures involve centralization.11 Three of the four stipulate some aspect of the cooperation problem the states are facing as the independent variable: Centralization increases with the severity of the enforcement problem, with uncertainty about behavior, and with uncertainty about the state of the world.12 The fourth conjecture is a transactions-cost argument: Centralization increases with number, where number can capture the literal number of states or their heterogeneity, or both.13
To elaborate a bit on the underlying logic of these conjectures, consider the following: Uncertainty about the behavior of states engaged in a cooperative endeavor can be reduced by developing centralized monitoring systems, whereas uncertainty about the state of the world can be partially offset by the pooling of information, which is often done most efficiently through delegation. When enforcement problems are present, some kind of delegated punishment (for example, through a body like the Security Council) can help deter the problem of cheating. States facing internal commitment problems, perhaps because of a tumultuous recent political history, can often more easily commit credibly to cooperative endeavors if they delegate some authority to a third party. Essentially, when cooperation problems are complex (here including enforcement problems and commitment problems as well as uncertainty about behavior and the state of the world), delegation helps states solve the problem or problems more efficiently.14 Regarding the transaction-costs conjecture, when heterogeneity is high, routine administration and decisionmaking is complicated among the participants; delegating some of those tasks can be efficient. Similarly, increasing the number of participants increases the transaction costs of implementing an international agreement; once again, delegation may be the efficient choice, allowing states to more easily coordinate their actions.
Rational Design does not distinguish between external and internal delegation, but according to its logic, one could argue that, if it is in states' interests to delegate, the larger and more heterogeneous the membership, the more likely it is that external delegation will be chosen. Externally delegating to a preexisting third party saves on the transaction costs of creating a new body and, more importantly, allows states to bypass the costly decisionmaking that is characteristic of large and heterogeneous groups. Additionally, external bodies [*pg 155] may be more objective and neutral, and hence they may be better equipped to come up with solutions that are acceptable to heterogeneous treaty members.
Many argue that nongovernmental organizations (NGOs) are increasingly important to international cooperation, particularly in the fields of the environment and human rights.15 It is worth considering explicitly what delegation to NGOs looks like, how it works, and what advantages it may provide to states. Rational Design suggests that states would choose to delegate functions to NGOs -- instead of to an internal body or an IGO (which is also composed of states) -- only if there is some explicit advantage to such delegation. Preliminary research on selective NGO participation appears to confirm this conjecture. "Nongovernmental organizations . . . are increasingly important participants in international environmental institutions," especially as these delegation agreements become more and more complex.16 Kal Raustiala has argued that the choice of states to give a role to NGOs is "based on the confluence of governmental incentives and NGO comparative advantages and resources."17 NGOs have even begun to take over what were traditionally "states-only" activities.18 Delegation to an NGO, particularly for functions like compliance monitoring and dispute resolution, may increase the impartiality of such processes and thus enhance the overall effectiveness of the agreement.19 NGOs may be able to serve this function through the links they create between local and global needs and actors.20 But they may also focus exclusively on compliance by their own governments or have very limited access to the information they need to make accurate and consistent judgments.21
|
III. DATA
The data used in this article are drawn from the United Nations Treaty Series.22 This database features the most comprehensive set of formal [*pg 156] international agreements concluded since the end of World War II as well as many agreements initiated in earlier periods, in particular those registered with the League of Nations. Unlike many databases that feature agreements from one country or region, which are either multilateral or bilateral, and which surround one issue area, like the environment, this database transcends the number and identity of parties, and the agreements it contains are divided into a number of issue areas. The random sample used here is conditional on four issue areas: economics, environment, human rights, and security. It includes both multilateral and bilateral agreements.23
Inclusion in the sample required that at least two state parties be involved, thereby eliminating any bilateral agreements between only one state and an international organization. With respect to Figure 1, "The Delegation Chain," found in the Bradley and Kelley article,24 the theories and data in this article are focused on delegations from the left-hand column to the middle column, although some agreements will mention redelegation from the collective body formed by the agreement to a third party. For example, in the American Convention on Human Rights, Pact of San José, Costa Rica (San José Pact),25 two internal bodies are created: the Inter-American Commission on Human Rights and the Inter-American Court of Human Rights. The tasks delegated to these bodies represent internal delegation. However, Article 40,26 which describes the role and functions of the Commission, mentions what Bradley and Kelley would call redelegation27 from the Commission to the preexisting Secretariat of the Organization of American States: the "Secretariat services for the Commission shall be furnished by the appropriate specialized unit of the General Secretariat of the Organization. This unit shall be provided with the resources required to accomplish the tasks assigned to it by the Commission."28 Hence this agreement includes both internal delegation and external delegation (accomplished through redelegation).
A coding instrument records the characteristics of the agreements. Importantly, the coders for this project are extensively trained in order to give them high levels of both competency and consistency, with the majority going through nine to twelve months of course-based training that includes both theoretical training and practice coding runs. Two coders independently code [*pg 157] each agreement using an online survey instrument. Upon completion, an intercoder-reliability report is generated for the 375 questions for which there are "quantitative" answers, for example, yes or no, multiple choice, or a specific number. The average coded agreement is characterized by disagreement on approximately fifteen questions, or four percent of the quantitative questions; the range so far has been between two percent and fifteen percent. The inconsistencies are then resolved through a close rereading of the agreement and supervised discussion involving the original coders, a trained graduate student, and the author.
One of the most extensive sections of the coding instrument is that which addresses the bodies (if any) created by the agreement. These bodies are usually composed of some kind of representative of the member states, but they can also be composed of experts chosen by the members. Because they are new bodies created by the agreement, they are a form of internal delegation. A shorter section of the coding instrument addresses tasks delegated to third parties and asks about the kind of tasks delegated as well as what kind of third party is implicated. This section captures external delegation. Hence, what the distinction is capturing is whether new bodies are created or existing ones are used.
In their concept article, Bradley and Kelley also include both internal and external delegation in their definition of international delegation.29 Within their framework, internal delegation would be delegation to a collective body or subgroup of states within the agreement (what some call a pooling of sovereignty), while external delegation would include delegation to a third party outside of the agreement.30 In Figure 1, internal delegation encompasses the first two rows of the middle column while external delegation includes the bottom row in Bradley and Kelley's article. In principle, each of the possible tasks delegated can be through either internal or external delegation. For example, in the data presented here, the secretariat function is sometimes handled through internal delegation and sometimes through external (and, of course, sometimes not delegated at all), as the first two columns of Table 1 illustrate.
To help bring these two concepts to life, consider the following agreement from the random sample. In an investment agreement between the United Kingdom and Egypt, the states must submit their dispute to an arbitral body if they cannot settle it diplomatically.31 The members of this body are chosen by the disputing states, but if they cannot find mutually acceptable members, the selection process is turned over to an external source, either the International Centre for the Settlement of Investment Disputes or the International Court of [*pg 158] Justice.32 Thus, the arbitral body is usually internal, but if the members cannot agree, it becomes external. The arbitral body listens to arguments, considers the facts, and, in this case, makes binding recommendations to resolve the dispute.33
Once a body has been identified as internal, the coders are asked to identify in detail the characteristics of the members, the procedures for making decisions, and other details about the functioning and purpose of the body, including what tasks are delegated to it. The list of possible delegated tasks appears in Table 1 below.34
|
IV. DESCRIPTIVE STATISTICS
This part presents a first look at the landscape of delegation across the random sample of agreements. Figure 1 shows the incidence of delegation and whether it is internal, external, or both.
One of the most striking things about Figure 1 is how widespread delegation is, with over half of the international agreements calling for it in some form or another. About the same number of agreements call for internal and external delegation, with the use of both in a single agreement being somewhat less common but certainly not negligible. This is not surprising, given that there are costs to creating bodies as well as to delegating to existing ones; hence, states may have an incentive to choose the form of delegation selectively. It is also notable that although external delegation is often cited as having higher "sovereignty costs," or as placing higher constraints on signatory states, at first glance the data above do not suggest that states are any less likely to use external delegation. This suggests that perhaps external delegation is not viewed by states in such terms, but instead is seen as being the more effective form of delegation in many cases. Of course, any firm conclusions must take into account just what gets delegated internally versus externally.
Although all of these third parties are preexisting intergovernmental organizations (IGOs), in six agreements in the sample, NGOs are mentioned as well. Given the field's interest in the role of nonstate actors, it is worth looking at what delegation to NGOs looks like in the random sample. Out of those six agreements that delegate functions to NGOs, four are in the area of human [*pg 160] rights. In the San José Pact, signatory states give an NGO the authority to bring a petition against a signatory state if the NGO believes that a violation has occurred.35 Specifically, the Pact calls on "[a]ny person or group of persons, or any nongovernmental entity legally recognized in one or more member states of the Organization" to "lodge petitions with the Commission containing denunciations or complaints of violation" of the Convention.36
Similarly, the Convention for the Protection of Human Rights and Fundamental Freedoms delegates compliance-monitoring functions to an NGO.37 Article 25 notes that
[t]he Commission may receive petitions addressed to the Secretary-General of the Council of Europe from any persons, nongovernmental organisation or group of individuals claiming to be the victim of a violation by one of the High Contracting Parties of the rights set forth in this Convention, provided that the High Contracting Party against which the complaint has been lodged has declared that it recognizes the competence of the Commission to receive such petitions.38
The agreement therefore gives NGOs the authority to monitor state compliance with the agreement and to bring petitions against noncompliant states when appropriate.
Figure 2 shows the incidence of delegation by issue area.
Cooperation without delegation is most common in the security and economic issue areas. For security agreements, this descriptive statistic seems to [*pg 161] confirm the conventional wisdom that states may be less willing to turn issues pertaining to national security over to an international body. With respect to economic agreements, an inverse relationship between precision and delegation can be expected, and economic agreements are, on average, far more precise than agreements in any of the other three issue areas. 39 Cooperation without delegation is least likely in human rights. Moreover, and interestingly, human-rights agreements never incorporate internal delegation solely; rather, they either delegate only to third parties or they create an internal body and delegate to a third as well. This is not surprising, given that human-rights agreements tend to govern internal affairs; hence, delegation to a third party may enhance credibility.
These simple, descriptive statistics open up a set of questions regarding the role of the existing international environment -- most importantly, existing institutions -- in the formation of new agreements and the process of delegation. The formation of new international law is often an iterative process, with states building off of old institutions to adapt to their new needs.
The types of delegated authority discussed by Bradley and Kelley can include functions that represent a substantial delegation of authority. These include, for example, legislative tasks such as the power to amend treaties and mandate compliance, adjudicative tasks such as the settlement of disputes, and compliance-monitoring and enforcement, such that the body to which power is delegated works to enforce the treaty either through centralized investigations or more locally through peer pressure. The delegated bodies are used for less-central tasks as well, for example, to create administrative rules to implement the treaty, set the agenda for the organization, and provide research and advisory support.40 Table 1 shows what kinds of tasks are actually delegated across the random sample of agreements. Although the list of tasks does not overlap entirely with Bradley and Kelley's, it includes many of the same types of functions and levels of delegation, including high levels of delegation, such as compliance monitoring, dispute settlement, and enforcement, as well as delegation of less central tasks such as administrative functions, the drafting of rules to ease implementation, and information dissemination.
Table 1: Tasks Delegated to Institutions
As the table indicates, dispute resolution is the most commonly delegated function, both internally and externally.41 External delegation of dispute resolution usually involves delegating authority to an existing arbitration tribunal or an international court.
For tasks other than dispute resolution, some facet of the implementation of the agreement, as well as the analysis and dissemination of information, are the tasks most often internally delegated. Tasks related to the monitoring and punishment of noncompliers are also at times delegated, suggesting that internal delegation may be an important enforcement tool as well. It is also worth noting that the task of making laws in addition to those stipulated in the agreement is internally delegated in about ten percent of the agreements and is never externally delegated. This strongly suggests that states want to retain control over the evolution of their cooperation and are not apt to delegate such tasks to a third party. Secretariat and administrative duties, collection of information, compliance monitoring, and punishment for noncompliance are the tasks that are most frequently externally delegated. In fact, ten percent of the agreements delegate monitoring functions to an external body. Significantly, the table shows that delegation is used for a wide range of tasks, tasks that are important and in some cases essential to the functioning of agreements. Indeed, delegated bodies are involved in the implementation, revision, and monitoring of the agreement and often have the power to settle disputes and enforce compliance. The actions of delegated bodies in these functions will have real effects on signatory states.
Given the statistic in Figure 1 -- that about thirty percent of the agreements use both internal and external delegation -- the third column of Table 1 delves into whether a redundancy is built into some agreements, or whether different tasks are delegated internally and externally.
Delegation of the same tasks to new and old institutions is not all that common, except in dispute resolution, when it occurs for about twenty percent of the agreements in the sample. It is likely that most of these cases are ones in which dispute resolution is first delegated to an internal body and then to an external body only if no agreement can be reached at the first stage. One example in which the same (nondispute resolution) functions are delegated to new and old institutions is the Treaty for the Prohibition of Nuclear Weapons in [*pg 165] Latin America (the Tlateloco Treaty).42 This agreement delegates compliance monitoring internally to the "Agency for the Prohibition of Nuclear Weapons in Latin America" (OPANAL), and externally to the International Atomic Energy Agency (IAEA).43 OPANAL is given the responsibility of holding "periodic . . . consultations" among member states on matters relating to member compliance and obligations.44 The IAEA is tasked with conducting special investigations of member-state activity when necessary or when requested by another signatory.45 Initially, both OPANAL and the IAEA were given this responsibility, but following a 1992 amendment, only the IAEA retains it. By delegating compliance monitoring to both internal and external bodies, states may have hoped to combine the state control provided by internal delegation with the "teeth" afforded by having an external body involved. What they learned, perhaps, is that, in this case, the external body provided the needed credibility.
The next issue, relationships between voting rules and internal delegation, is significant because it relates to the autonomy of the international body as well as to the costs states face when they delegate authority. The autonomy of a body, Bradley and Kelley suggest, is determined by the level of state oversight, including the reporting requirements of the body, the precision of the body's mandate, the decisionmaking procedures of the body (for which unanimous processes reduce the body's independence), the permanence of the body, and the mechanism through which the body gets its funding.46 The analysis of voting rules here captures the decisionmaking-procedure aspect of Bradley and Kelley's autonomy concept. Bodies that have voting rules requiring unanimity ("unanimity rules" or "unanimity voting") are likely to have lower levels of autonomy than those with some type of majoritarian voting rules. Similarly, bodies that require supermajorities may have less autonomy than those that need only simple majorities. The level of autonomy given to the international body, Bradley and Kelley note, will also affect the cost of delegating authority.47 When states continue to control the body to which power is delegated, the cost of delegation remains low and they are not forced to compromise on their preferred policy all that much.48 However, when the body is given substantial independence, states sacrifice more of their individual autonomy through the delegation of authority and may be forced to accept policies that vary drastically from their ideal.49 By considering how voting power is distributed, the analysis of voting rules here sheds light on the costs of delegation, probing not only [*pg 166] whether states rely more heavily on unanimity or majority voting in their delegation, but also whether weighted or one-vote-per-member voting rules are used in those cases where majority voting is permitted. In other words, do the rich and powerful control the delegation more?
It may be also be the case that unanimity voting rules are more common in instances of delegation in specific issue areas -- for example, security -- or when the delegation includes certain types of functions -- for example, imposing sanctions or amending the treaty. If powerful states rely on weighted voting to control their delegation, not only will the delegated body have less independence overall, but the autonomy costs50 may also be lower for these powerful states than for smaller signatory nations. More specifically, we can see whether agreements involving superpowers are more likely to have asymmetric voting rules for any internal bodies. If this is the case, it would suggest that powerful countries do carefully protect their autonomy when allowing delegation. Findings suggesting that states do guard their delegation would not necessarily imply delegated bodies are weak, but rather that state actors use delegation as a tool to advance their interests but are careful not to let it get out of hand.
Table 2: Voting Rules for Internal Delegation
Table 3: Determination of Representation for Internal Delegation
Table 4: Allocation of Votes for Internal Delegation
The tables above indicate, first, that the large majority of internally delegated bodies have a fixed number of representatives per nation state and voting rules that allow one vote per member state, and, second, that unanimity voting rules are rare (less than ten percent of the bodies). This suggests that such bodies are not dominated by certain states (for example, large powers) and that, at least for internal delegation, states do not often guard their delegation with unanimity or weighted voting. These descriptive statistics support the argument that delegated bodies are given the autonomy to take real action and thus can have real effects on signing states.
Further, probes into the data suggest that there is a negative relationship between the inclusion of a unanimity rule and the agreement issue areas of security, environment, and human rights. The results do not approach statistical [*pg 168] significance, however. In economic issue areas, there is a positive relationship with the inclusion of a unanimity rule, a result that is significant at the 0.10 level. For types of functions that are delegated to internal bodies, there is a positive relationship between the delegation of dispute-resolution functions and the use of a unanimity rule. This result is not statistically significant at conventional levels (p=0.1488). There is also an inverse relationship between the inclusion of unanimity rules and the delegation of amendment functions as well those of overseeing complaints and punishments, but they are not statistically significant.
Finally, the relationship between whether a superpower is a member and the use of weighted voting depends on how weighted voting is coded. When, in the table, other types of voting-allocation rules are coded as weighted voting, the relationship between weighted voting and superpower membership in the agreement is statistically significant and positive. This suggests that when either the United States or (the former) USSR is a signatory to an agreement, it is likely to seek out weighted voting rules. (However, when other types of voting- allocation rules are coded as either not weighted or as missing data, the relationship between the presence of a superpower party to the treaty and weighted voting is not statistically significant. This may be a result of a very small group of agreements that have any weighted voting, by any coding scheme. When other is coded as weighted, only three agreements create bodies with weighted voting; when other is either coded as not weighted or as missing data, the number of agreements with weighted voting is only one.)
|
V. FINDINGS: WHEN IS DELEGATION MOST LIKELY?
Rational Design theory calls attention to a set of independent variables whose presence would increase the likelihood that states incorporate some kind of delegation into their agreements.51 The subsections that follow address each of these variables. Because these data are new, I discuss the probits one variable at a time before turning to the multivariate analysis below.
A. Delegation and the Complexity of the Cooperation Problem
Dispute-resolution provisions are much more common in agreements that deal with complex cooperation problems.52 The explanation for this variation in dispute-resolution provisions based on Rational Design theory is that states rationally include dispute-resolution provisions only when they expect that such provisions will be needed in the future. (Recall that each of the problems included in the variable complex cooperation problem calls for centralization to help solve the problem.53) In signing agreements with more complex [*pg 169] cooperation problems, states are likely to anticipate future disagreements, and rightly so, and more likely to see dispute-resolution provisions as necessary. This explanation is strongly supported in the statistical analyses presented here.
When the delegation is broadened to include all forms of delegation, not just dispute resolution, both external and internal delegation appear to be more common when there are complex cooperation problems (see Table 5). When states are forced to overcome complex cooperation problems, they use delegation as a tool to accomplish this challenge. Delegation can allow them to deal with compliance monitoring, dispute resolution, and even the sanctioning of defecting states, thereby addressing the enforcement problem. It can also be used to address the commitment problem by increasing the extent to which the agreement "ties the hands" of signing states. The effect is larger for external delegation, perhaps because external organizations are seen as more neutral and credible than internal bodies,54 which may increase their effectiveness in overcoming cooperation problems. This would be the case especially when states are trying to solve commitment problems: a third party would be seen as more credible than the state itself when the state has voting power. (In terms of marginal effects, the presence of a complex problem increases the probability of external delegation by nearly forty percent and internal delegation by twenty-one percent).
Table 5: Results of Probit Analysis of Complexity and Delegation
B. Delegation and Number and Heterogeneity of Parties
Rational Design theory predicts that delegation will increase with the number of states involved in a cooperative endeavor.55 Using the number of participants at the original negotiation of the agreement as a measure, the effect [*pg 170] number has on the choice of delegation can be analyzed by performing a probit regression. Table 6 illustrates the results.56
Table 6: Results of Probit Analysis of Number of Parties and Delegation
Table 6 illustrates a positive and significant relationship between number and external delegation only. With respect to marginal effects, as the number of parties changes from two to three, the probability of external delegation increases by eight percent, whereas the probability of internal delegation increases only by 1.2%.57
According to the theory articulated above, delegation should increase with the heterogeneity of the parties.58 The narrower argument is that delegation will be used increasingly to solve disputes as states' preferences become more diverse. This prediction constrasts directly with that by Hawkins et al., which states, "The greater the preference heterogeneity of any group of states, therefore, the less likely they will be to delegate to an [international organization]."59
Table 7 shows some simple probit results from the perspective of the relationship between the heterogeneity of members and delegation.60
Table 7: Results of Probit Analysis of Heterogeneity of Members and Delegation
As Table 7 illustrates, there does seem to be a positive relationship between heterogeneity and both kinds of delegation, with the effect on external delegation being larger and more significant, as predicted in the theory section above. Substantively, a change in the heterogeneity of parties from a middle level of zero to a high level of one increases the probability of external delegation by nearly fifty percent; the same change in heterogeneity increases the probability of internal delegation by only twelve percent.
When both heterogeneity and number of parties are included in the probit analysis, as in Table 8, the significance of both of these variables for external delegation remains high. This implies that both heterogeneity and number of parties have independent effects on the choice of external delegation.
Table 8: Results of Probit Analysis of Heterogeneity and Number of Parties and Delegation
Substantively, when controlling for the number of parties, the effect of a change in the heterogeneity of members from zero to one on the probability of external delegation decreases slightly to thirty-one percent, while it leaves the marginal effect of heterogeneity on the probability of internal delegation nearly [*pg 172] unchanged. In addition, controlling for heterogeneity slightly reduces the effect of the number of parties on the probability of external delegation to 6.5%, while the number of parties, controlling for heterogeneity, does not affect the probability of internal delegation.61
The results in Tables 6-8 confirm the prediction that larger and more heterogeneous groups of states are more likely to choose external rather than internal delegation.
C. Other Possible Explanatory Variables
Other explanatory variables (democracy, the presence of a superpower, and the member states' level of risk aversion), while not stemming from Rational Design, are common control variables in empirical analyses of international relations.62 Democracy is measured by the mean polity score of treaty signatories.63 As Table 9 illustrates, both forms of delegation are less common when the parties are democracies than when they are not.
Table 9: Results of Probit Analysis of Democracy and Delegation
More concretely, as the average level of democracy among parties to an agreement changes from zero (representing a relatively heterogeneous group of members) to ten (representing an agreement in which all the members are democracies), the probability of external delegation decreases by thirty-seven percent and the probability of internal delegation decreases by twenty-five percent.
What might explain this striking finding? One possible explanation is that these findings may relate to those on heterogeneity: when more signing states [*pg 173] are democracies, perhaps they are also more homogeneous, with similar preferences and similar interests. One way of investigating this possibility is to look at the correlation coefficient between the two variables, democracy and heterogeneity. It turns out that the correlation is -0.3552, with a p-value of 0.006 (and thus statistically significant); hence democracies are quite homogeneous, at least in terms of the measure used here, which is similar voting patterns in the United Nations General Assembly. Still, there are other possible explanations. It may be that democracies are more able to use informal mechanisms in their cooperation. Finally, perhaps democracies worry more than nondemocracies about the democratic values that international delegations might compromise -- for example, participation, accountability, and responsiveness.64 In any event, more in-depth research is needed to explain this finding.
Considered next is whether the presence of a superpower among the signatories affects the probability of having external or internal delegation.65 A superpower may be reluctant to delegate authority and reduce its autonomy. Yet a superpower signatory does not significantly change the probability of either type of delegation. In addition, the substantive effect of a superpower signatory is very small for both external and internal delegation. Having the United States or another superpower like the former Soviet Union as a signatory increases the probability of external delegation by two percent, while it reduces the probability of internal delegation by ten percent.
Finally, anarchy could make risk-averse states cautious about delegating, especially delegating important tasks such as monitoring, implementation, and dispute resolution. As early as James Madison, those writing about delegation have demonstrated that delegated power can be used against the principal delegating it.66 Of course, the development of delegation theory has focused on how contracts can be structured to minimize risks like agency slippage.67 The [*pg 174] bottom line, however, is that tradeoffs are inherently involved.68 The traditional view, in summary, is that "[t]he potential for inferior outcomes, loss of authority, and diminution of sovereignty makes states reluctant to accept hard legalization -- especially when it includes significant levels of delegation."69 Thus, states facing enforcement problems are less likely to include delegated dispute-resolution provisions in their international agreements. Moreover, since external delegation may be harder for states to control, and thus imply more risk, this effect is likely to be larger for external delegation than for internal delegation.
The four different probit analyses test the relationship between risk aversion and delegation overall as well as the delegation of monitoring, implementation, and dispute resolution.70 Risk aversion and delegation do not appear to be significantly related, except in the case of the internal delegation of dispute-resolution provisions, suggesting that states seek the benefits they can derive from delegation despite the risks involved. In other words, states are not single-mindedly avoiding risk as many Realists would predict; rather, they make tradeoffs based on cost-benefit calculations. Still, a more fine-grained analysis is called for once a greater data set allows it. For example, among other things, one would need to look into whether states can opt out of the delegated dispute resolution by adding reservations to the agreements.71
D. Relationship Between Delegation and Other Design Provisions
Examining the relationship between delegation and other design provisions logically begins by describing other work that has been pursued on the relationship between delegation of dispute resolution and the degree of precision of an agreement's main substantive goals.
1. Delegation and Precision
Delegation of dispute-resolution authority in an international agreement should be inversely related to the degree of precision accorded to the agreement's substantive goals.72 A bivariate probit model indeed reveals a statistically significant inverse relationship between these two design variables. Moreover, consistent with the results above, the variable complex cooperation problem has a positive and significant effect on dispute-resolution delegation provisions. 73
2. Delegation and Withdrawal Clauses
There are several hypotheses for why the use of delegation may be associated with the use of withdrawal clauses. In their article, Bradley and Kelley capture the notion of withdrawal clauses by considering the permanence of the delegated body and whether it is easy or difficult for states to remove themselves from the jurisdiction of the delegated body or the associated agreement, or both.74 They suggest that even the existence of withdrawal clauses might not remove the practical difficulty of removing oneself from an agreement or from part of an agreement, especially if one agreement is embedded within others as is the case for the European Convention on Rights and the Council of Europe.75
On the other hand, in her paper for this symposium, Hathaway argues that withdrawal clauses may serve to protect states against time-inconsistent preferences -- that is, cases in which state preferences change over time.76 Such clauses allow states to remove themselves from forms of delegation that they find increasingly counter to their revised interests. This is very similar to the Koremenos argument that "[w]ithdrawal clauses are responses to shocks that alter a state's basic interest in cooperation."77 These shocks are rare, but the risk they impose is great. Hence withdrawal clauses are pervasive, but not used very frequently.78 In this sense, withdrawal clauses play a role similar to that of provisions that allow states to revoke authority from a delegated body or to remove themselves from its jurisdiction. They provide flexibility and preserve a [*pg 176] degree of state independence. As a result, one would expect to see withdrawal clauses in treaties that also include delegation, particularly external delegation.
Table 10 illustrates a significant and positive correlation between delegation provisions and withdrawal clauses. What this suggests is that delegation is not trivial for states; rather, it impedes their autonomy. Hence, agreements that call for delegation, in particular external delegation, are far more likely to include withdrawal provisions than those that do not.
Table 10: Correlation Between Withdrawal Clause and Delegation
3. Delegation and Finite Duration Provisions
Duration provisions provide a form of insurance against shocks that may influence how an agreement affects state interests.79 Because external delegation may give some real power to an external body, states may perceive a higher degree of uncertainty regarding the possible outcomes of such an agreement that includes delegation. As a result, signing parties may be more interested in including a finite duration clause as protection against a shock due to some unexpected action by the delegated body. Although this may be true of internal delegation to some extent, it is more likely with external delegation. On the other hand, bodies, whether external or internal, may be able to provide the kind of flexibility desired by states when faced with such uncertainty. The bodies can be delegated the authority to adjust the terms of cooperation in response to shocks. If this is the case, delegation could be a substitute for finite duration provisions.
The data in Table 11 provide an insight into which of these possible explanations seems more supported.
Table 11: Correlation Between Finite Duration and Delegation
The results suggest that agreements that include external delegation are less likely to have finite duration, supporting the latter argument that these bodies actually provide some desired flexibility. The correlation between internal delegation and finite duration is nearly zero and is statistically insignificant, leading to the conclusion that there is no relationship between the two.
E. Multivariate Analysis
As Tables 12 and 13 demonstrate, in the multivariate regression, all the variables explaining external delegation remain correctly signed, and only heterogeneity loses statistical significance across both dependent variables.80 This is not surprising, since the sample size is relatively small. Controlling for heterogeneity, complexity, and democracy, the number of parties still predicts more external delegation. Substantively, a change from two to three parties increases the probability of external delegation by seven percent. This lends additional support to the argument that parties use delegation to reduce transaction costs, giving more confidence to the Rational Design conjecture that centralization increases with number. The effect of the number of parties on internal delegation, although negative, is very small and statistically insignificant, all other things being equal. If the underlying cooperation problem is complex, the probability of external delegation increases by thirty-six percent, and the probability of internal delegation increases by seventeen percent, holding everything else constant. This result is very strong with respect to external delegation, lending strong support to the Rational Design hypotheses linking complex cooperation problems to centralization. The change in democracy lowers the probability of both external and internal delegation by about thirty-two percent, holding everything else constant.
Table 12: Multivariate Analysis
Table 13: Marginal Effects: Multivariate Results
|
VI. CONCLUSION
This article takes a first, systematic look at the incidence of delegation across a random sample of agreements. Not only is delegation widespread; these results also confirm the observable implications of the Rational Design [*pg 179] project, which were deduced and articulated five years earlier -- long before any of the current data had been collected. This article is therefore a contribution to a broader research agenda on rational institutional design as well as on international delegation.
A number of interesting empirical findings stand out. For example, states delegate the least in the security issue area, but when they do delegate, they tend to use external delegation almost twice as much as internal delegation. More theoretical and empirical work needs to be done to explain such findings. For instance, do state representatives in security agreements feel that when they must delegate, they want a more neutral third party to lessen the risk of any conflict escalation? Another striking finding is how little states use unanimity or even supermajority voting rules to protect themselves or to limit delegated authority.
Additionally, the findings on heterogeneity leading to less delegation contrast with the predictions in the work of Hawkins et al.81 The predictions presented here are theoretically based on the work of Erhlich and Posner, who focus on dispute resolution.82 The findings are based on data concerning all forms of delegation, including the delegation of dispute resolution, and across four issue areas, whereas the work of Hawkins et al. focuses on a particular kind of delegation: the implementation of policy.
The analyses in this article all focus on the internal-external delegation distinction. This distinction is warranted, given the statistically significant differences between the two types of delegation in the analyses presented here. Bradley and Kelley suggest that there are significant differences among international delegations beyond the internal-external distinction.83 Empirically, with a larger number of agreements in the sample, researchers can begin to break down the analyses and probe what kind of delegation is present in each issue area as well as investigate the voting rules and the particular characteristics of the states involved.
Another interesting analysis that would be possible with a larger data set is to match specific cooperation problems with various kinds of delegation, including the external-internal distinction, and to probe whether the prediction in Hawkins et al.84 on the delegation of policy implementation holds up across a larger set of agreements. The initial results in this article show that such future research is promising and should be conducted.
Additionally, as more data become available, multivariate and interactive analyses should be employed to achieve not only a greater level of certainty about the relative importance of delegation for different cooperation problems, but also a more sophisticated understanding of under what conditions which particular tasks will be delegated.
Finally, increasing the sample will also allow a time-series analysis. This is not trivial, given that one set of questions we can then study will be the effects of the bipolar world on delegation and membership in those agreements that call for delegation. For instance, the bipolar world may have inhibited widespread and deep delegation to global organizations; on the other hand, perhaps the bodies that were created during the Cold War were narrow in terms of membership but called for substantial delegation. So, polarity (whether the world be bipolar as it used to be or unipolar as it is now) may affect the scope and depth of delegation as well as the nature of problems resolved through international institutions. Another issue that could then be addressed is whether states increased the amount or depth of delegation to NGOs.
These are significant questions that need to be addressed, and the expansion of datasets like the one used in this paper will help uncover new puzzles and provide convincing empirical evidence about the changes occurring in the international environment as a result of the rise in nonstate actors as well as of the demise of the Cold War.
APPENDIX: TABLES OF AGREEMENTS IN THE RANDOM SAMPLE ORGANIZED BY SUBJECT
[*pg 187]
[*pg 188]
[*pg 189]
[*pg 190]
[*pg 191]
[*pg 192]
|
FOOTNOTES
Copyright © 2008 by Barbara Koremenos.
| * | Associate Professor of Political Science, University of Michigan. The editors, the other authors featured in this special issue, and George Tsebelis made many useful comments on earlier drafts, greatly improving this article. Michelle Allendoerfer, Eleni Gouvas, Jennifer Kavanagh, Johannes Urpelainen, and Peter Wennerholm provided excellent research assistance. This article could not have been completed without them. |
| 1. | See generally Curtis A. Bradley & Judith Kelley, The Concept of International Delegation, 71 LAW & CONTEMP. PROBS. 1 (Winter 2008). |
| 2. | See, e.g., Michael J. Tierney, Delegation Success and Policy Failure: Collective Delegation and the Search for Iraqi Weapons of Mass Destruction, 71 LAW & CONTEMP. PROBS. 283 (Winter 2008). |
| 3. | Individual scholars know a great deal about the nature and consequences of delegation relationships within certain international agreements and organizations. See, e.g., Alexander Thompson, Coercion Through IOs: The Security Council and the Logic of Information Transmission, 60 INT'L ORG. 1 (2006) (examining the United Nations Security Council to explain why states channel coercive policies through international organizations). For such consequences within certain issue areas, see, e.g., Lisa Martin, Distribution, Information, and Delegation to International Organizations: The Case of IMF Conditionality, in DELEGATION AND AGENCY IN INTERNATIONAL ORGANIZATIONS (Darren G. Hawkins et al. eds., 2006); Karen J. Alter, Private Litigants and the New International Courts, 39 COMP. POL. STUD. 22 (2006) (showing a trend towards creating and using international courts); David L. Nielson & Michael J. Tierney, Delegation to International Organizations: Agency Theory and World Bank Environmental Reform, 57 INT'L ORG. 241 (2003) (examining lending policies for environmentally significant projects administered through the International Monetary Fund and the World Bank). All of these scholars have gathered their own very detailed and compelling evidence within these agreements or organizations or within issue areas. Nonetheless, generalizing from this very small sample of institutionalized international cooperation and international delegation is often misleading. |
| 4. | See generally Barbara Koremenos et al., The Rational Design of International Institutions, 55 INT'L ORG. 761 (2001) (elaborating the theory of Rational Design as applied to international organizations). |
| 5. | Included in institutionalized international cooperation are any explicit arrangements, negotiated among international actors, that prescribe, proscribe, or authorize behavior; thus this includes the over 50,000 international agreements that are registered with the United Nations. |
| 6. | Koremenos et al., supra note 4, at 776. |
| 7. | Id. at 775. |
| 8. | Barbara Koremenos, THE CONTINENT OF INTERNATIONAL LAW (2007) [hereinafter KOREMENOS, CONTINENT]. In CONTINENT, I add the following possible cooperation problems: encouraging positive externalities, discouraging negative externalities, deadlock, and "other." The other category includes areas of cooperation such as foreign aid for which there are no or few strategic considerations and pure coordination games without uncertainty. More than one answer can be chosen for each agreement. In CONTINENT and in Barbara Koremenos, If Only Half of International Agreements Have Dispute Resolution Provisions, Which Half Needs Explaining?, 36 J. LEGAL STUD. 189 (2007) [hereinafter Koremenos, Which Half?], multiple examples of the operationalization of these cooperation problems are given, as well as elaboration on how they were coded. |
| 9. | Given that all the agreements in the sample include at least two state actors, whenever tasks are performed by a single focal actor, the states are delegating. |
| 10. | In univariate analyses, each variable in a data set is analyzed separately. |
| 11. | Koremenos et al., supra note 4, at 787-91. |
| 12. | Id. |
| 13. | Id. Specifically, the variable number can capture asymmetries with respect to both power and interests. In the empirical analyses below, number, power, and heterogeneity of interests are included as separate variables. |
| 14. | Cooperation problems not included in the category of complex problems include encouraging endeavors with positive externalities, discouraging those with negative externalities, and solving coordination games. Though uncertainty about preferences could be considered a complex problem, few if any of the agreements in the sample have that as one of their underlying problems, which is interesting in and of itself. See Koremenos, CONTINENT, supra note 8. |
| 15. | See, e.g., Margaret E. Keck & Kathryn Sikkink, ACTIVISTS BEYOND BORDERS: ADVOCACY NETWORKS IN INTERNATIONAL POLITICS passim (1998) (arguing that NGOs have made an impact in human rights and environmental politics by bringing attention to particular issues and changing states' perceptions of their national interests). |
| 16. | Kal Raustiala, States, NGOs, and International Environmental Institutions, 41 INT'L STUD. Q. 719, 732-33 (1997). |
| 17. | Id. at 720. |
| 18. | Id. at 719. |
| 19. | Id. at 728-30. |
| 20. | Thomas Princen & Matthias Finger, ENVIRONMENTAL NGOS IN WORLD POLITICS 33 (1994). |
| 21. | Oliver Meier & Clare Tenner, Nongovernemental Monitoring of International Agreements, in VERIFICATION YEARBOOK 207, 217 (Trevor Findlay ed., 2000). |
| 22. | United Nations Treaty Collection, http://untreaty.un.org/English/access.asp (last visited Jan. 28, 2008). The data are gathered as part of a research project supported by the National Science Foundation CAREER Award: "Designing International Agreements: Theoretical Development, Data Collection, and Empirical Analysis" (SES-0094376). The Internet collection at the time the sample was drawn contained over 34,000 international agreements "which have been published in hard copy in over 1,450 volumes, which corresponds to all treaties and subsequent actions registered up to December 1986." Id. |
| 23. | In the sample used for this article, sixty-four of the ninety-seven agreements are bilateral. The reason for this is that bilateral cooperation is far more prevalent than multilateral cooperation. For example, at the time the sample was drawn, the online version of the UNTS contained 2,330 multilateral agreements and 32,936 bilateral agreements. Nothing about the design of the data-collection project precludes it from including other issue areas; the four chosen are included because they are currently the most important in the international cooperation literature. See the Appendix for a list of the set of agreements used in the empirical analyses. |
| 24. | Bradley & Kelley, supra note 1, at 6. |
| 25. | Organization of American States, American Convention on Human Rights, Pact of San José, Costa Rica, Nov. 22, 1969, O.A.S.T.S. No. 36, 1144 U.N.T.S. 144 [hereinafter San José Pact]. |
| 26. | Id. at 154. |
| 27. | Bradley & Kelley, supra note 1, at 6. |
| 28. | San José Pact, supra note 25, at 154. |
| 29. | Bradley & Kelley, supra note 1, at 3-9. |
| 30. | Id. at 6. |
| 31. | Agreement for the Promotion and Protection of Investments, U.K.-Egypt, June 11, 1975, 1032 U.N.T.S. 32. |
| 32. | Id. |
| 33. | This is one of the most "controversial cases" of internal delegation because it could be claimed that, when disputing members must each choose an arbitrator to form a panel, these arbitrators are neutral third parties. Nonetheless, they could just as easily be retired judges from the member states and hence biased. The conceptualization of internal delegation in this hardest case stems from the assumption that the units of analysis that bargain or negotiate at the international level are the states-not the governments within those states. The unitary-actor assumption is a good baseline as well as a focal point. Although it is known that domestic politics implies that it is not always the case that the state should be modeled as unitary, it becomes very difficult in reading agreements to ascertain which level of the state is the one bargaining, influencing, being influenced, et cetera. There is simply not enough detail in agreements to discover such things. In contrast, the state is a salient unit. Given this, whenever a body is (or is likely to be) composed of some representative of at least some of the member states, this body is coded as internal-that is, a body created by the agreement and, hence, a form of internal delegation. For example, suppose that when Bolivia and the United States need to arbitrate a dispute, two very professional lawyers, one from the United States and one from Bolivia, with international reputations for being fair, are appointed as the arbitrators. If the government were used as the unit of analysis, this would be external delegation because the lawyers do not represent the governments. But given the assumption that the nation-state is the unit of analysis, the lawyers are citizens of the disputing parties as well as experts. Hence, this is a case of internal delegation. Other than the delegation of certain dispute resolution, all other forms of internal delegation are clear-cut examples of collective bodies or subgroups of states. |
| 34 | . See Barabara Koremenos, Glossary: Delegated Tasks, http://sitemaker.umich.edu/koremenos/files/glossary_-_delegated_tasks_jan08.doc (last visited Feb. 6, 2008) for a definition and example of each of the tasks listed in Table 1. |
| 35. | San José Pact, supra note 25, at 155. |
| 36. | Id. |
| 37. | Convention for the Protection of Human Rights and Fundamental Freedoms, Nov. 4, 1950, 213 U.N.T.S. 222. |
| 38. | Id. at 236-38. |
| 39. | Barbara Koremenos, AN ECONOMIC ANALYSIS OF INTERNATIONAL RULEMAKING (2007). |
| 40. | Bradley & Kelley, supra note 1, at 10-16. |
| 41. | In her contribution to this volume, Karen Alter elaborates four roles that a judiciary can take: dispute resolution, administrative review, enforcement, and constitutional review. Karen Alter, Delegating to International Courts: Self-Binding vs. Other-Binding Delegation, 71 LAW & CONTEMP. PROBS. 37 (Winter 2008). Thus far in the random sample, only one agreement has actually created a court: The San José Pact created the Inter-American Court of Human Rights, which has constitutional review and dispute resolution capabilities (adjudication). The court does not have enforcement capabilities; it can only make recommendations on "binding" rulings (rulings are only binding if member states recognize the jurisdiction of the court). Therefore, with that one exception, whenever a court, such as the ones Alter analyzes, is mentioned in the agreement, it is called upon to help with dispute resolution between the parties to some other agreement, with none of the other functions being mentioned. |
| 42. | Treaty for the Prohibition of Nuclear Weapons in Latin America, Feb. 14, 1967, 634 U.N.T.S. 326. |
| 43. | Id. at 334, 340. |
| 44. | Id. at 334. |
| 45. | Id. at 342-44. |
| 46. | Bradley & Kelley, supra note 1, at 20. |
| 47. | Id. at 28-30. |
| 48. | Id. at 30 |
| 49. | Id. at 28. |
| 50. | Autonomy costs are defined by Bradley and Kelley. Id. |
| 51. | See supra Part I. |
| 52. | Koremenos, Which Half, supra note 8, at 189. |
| 53. | See supra Part II. |
| 54. | See Kenneth Abbott & Duncan Snidal, Why States Act Through Formal International Organizations, 42 J. CONFLICT RESOL. 3, 16 (1998). |
| 55. | See supra Part II. |
| 56. | A log of the number of participants is used, since the unlogged variable is highly right-skewed and its log is almost perfectly normally distributed. |
| 57. | I selected two parties since sixty-four of the treaties are bilateral and hence a one-unit increase is standard. Then I took the natural log of 2 and 3 to perform the calculations. |
| 58. | See Isaac Ehrlich & Richard Posner, An Economic Analysis of Rulemaking, 3 J. LEGAL STUD. 257, 273 (1974) ("For example, as suggested earlier, we expect (and observe) more delegation of legislative power by the United States Congress than by the British Parliament because the costs of producing legislation are lower under the parliamentary system with its well disciplined parties and its effectively unicameral legislature."). |
| 59. | Darren G. Hawkins et al., Delegation Under Anarchy: States, International Organizations, and Principal-Agent Theory, in DELEGATION AND AGENCY IN INTERNATIONAL ORGANIZATIONS 3, 21 (Darren G. Hawkins et al. eds., 2006). |
| 60. | Gartzke and Jo's "Affinity of Nations Index," which measures the similarity in states' preferences based on voting preferences in the United Nations General Assembly, is used as a measure of heterogeneity. Eric Gartzke & Dong-Joon Jo, The Affinity of Nations Index, 1946-1996, Nov. 8, 2002, htttp://dss.ucsd.edu/~egartzke. Because the Affinity data are dyadic, I simply take the Affinity value for each bilateral agreement. For the multilateral agreements, I first create a dyad for each pair of signatories. Hence, if there are three signatories, there are three dyads; if there are four signatories, there are six dyads, and so on. For each multilateral agreement, the "weakest link assumption" is used, taking the Affinity value of the dyad with the least similar interests. |
| 61. | Marginal effect of heterogeneity is calculated holding number of parties (logged) at its mean (1.437) and the marginal effect of the number of parties is found holding heterogeneity at its mean (-0.309). |
| 62. | It could be argued that the presence of a superpower partly captures the asymmetry of power that is part of the variable, number, in rational design. Nonetheless, this is an imperfect measure at best. Some quantification of the power of the other signatories would be needed so that a measure of the discrepancy in power among agreement members could be created. |
| 63. | "Polity" is one of the most widely used measures of democracy. Among the dimensions of democracy captured in Polity are chief-executive constraints and the competitiveness of political participation. |
| 64. | See Neil Siegel, International Delegations and the Values of Federalism, 71 LAW & CONTEMP. PROBS. 93 (Winter 2008). |
| 65. | The superpower variable is equal to one if the U.S. or the USSR was a signatory and zero otherwise. |
| 66. | THE FEDERALIST NO. 62 (James Madison) (arguing that the legislative process is constructed to include an executive and a bicameral legislative branch in order to limit "excess law-making" which can result through delegation). His point is that delegation can be dangerous when it makes policy change and the construction of new laws too easy since this can lead to an abuse of power. |
| 67. | The literature has dealt with the issue of constraining agent "slippage" by emphasizing that the preferences of the agent and the principal are seldom aligned. OLIVER E. WILLIAMSON, THE ECONOMIC INSTITUTIONS OF CAPITALISM: FIRMS, MARKETS, RELATIONAL CONTRACTING 30 (1985). The principal may, therefore, choose or construct an agent that has similar preferences. DARREN G. HAWKINS & WADE JACOBY, How Agents Matter, in DELEGATION AND AGENCY IN INTERNATIONAL ORGANIZATIONS 199-228 (David A. Lake et al. eds., 2006). Principals may also do this by limiting the powers or mandate of the institution or by restricting the specialization of the agent to limit opportunism. EIRIK G. FURUBOTN & RUDOLPH RICHTER, INSTITUTIONS AND ECONOMIC THEORY: THE CONTRIBUTION OF THE NEW INSTITUTIONAL ECONOMICS 265-81 (2000); RODERICK D. KIEWIET & MATHEW D. MCCUBBINS, THE LOGIC OF DELEGATION 25 (1991). |
| 68. | For classic statements reflecting this view, see generally Armen A. Alchian & Harold Demsetz, Production, Information Costs, and Economic Organization, 62 AM. ECON. REV. 775, 775-95 (1972); Eugene Fama, Agency Problems and the Theory of the Firm, 88 J. POL. ECON. 288, 288-307 (1980). |
| 69. | Kenneth W. Abbott & Duncan Snidal, Hard and Soft Law in International Governance, 54 INT'L ORG. 421, 437 (2000). |
| 70. | All of these analyses use Bueno de Mesquita's risk-attitude measure, which uses a state's alliance portfolio to determine its level of risk aversion. Bruce Bueno de Mesquita, The War Trap Revisited: A Revised Expected Utility Model, 79 AM. POL. SCI. REV. 156, 158-59 (1985). Specifically, the closer a state is to having an alliance portfolio that maximizes its security, the more risk-averse it is presumed to be. In EUGene (Expected Utility Generation and Data Management Program), this risk attitude variable ranges from -1 (very risk-averse) to +1 (very risk-acceptant) and is based on region. See generally Scott D. Bennett & Allan C. Stam, A Universal Test of an Expected Utility Theory of War, 44 INT'L STUD. Q. 451 (2000). To measure risk attitudes in bilateral treaties, I calculate each signatory's risk attitude towards the other's region. For multilateral treaties, I find the mean of each signatory's regional risk scores. I use a "weakest link" assumption, and measure the agreement as a function of the risk attitude of the most risk-averse signatory. Finally, I invert the scale to create a measure of risk-aversion with -1 being the least risk-averse and +1 being the most risk-averse. |
| 71. | In fact, whether states are prohibited or not from adding reservations is another facet of institutional design that deserves further study. |
| 72. | Koremenos, supra note 39, passim (drawing on Ehrlich & Posner, supra note 58). This challenges the special issue on legalization, Legalization and World Politics, which does not suggest that these two design dimensions are substitutes. See Judith Goldstein et al., Legalization and World Politics, 54 INT'L ORG. 385 (2000). |
| 73. | This is consistent with Koremenos, Which Half, supra note 8, at 207-09. Ehrlich & Posner, supra note 58, also predict that greater heterogeneity across parties will lead to less precise contracts and therefore greater delegation of dispute resolution authority. |
| 74. | Bradley & Kelley, supra note 1, at 24. |
| 75. | Id. |
| 76. | Oona A. Hathaway, International Delegation and State Sovereignty, 71 LAW & CONTEMP. PROBS. 115 (Winter 2008). |
| 77. | Barbara Koremenos, Contracting Around International Uncertainty, 99 AM. POL. SCI. REV. 549, 561 (2005). |
| 78. | Id. |
| 79. | Id. at 549. |
| 80. | The correlation between number (logged participants) and heterogeneity is 0.6426†, whereas between democracy (Polity mean) and number (logged participants) it is -0.3363†. |
| 81. | Hawkins et al., supra note 59, at 21. |
| 82. | See generally Ehrlich & Posner, supra note 58. |
| 83. | See Bradley & Kelley, supra note 1, at 1-2. |
| 84. | Hawkins et al., supra note 59, at 21. |