This volume examines and exemplifies the usefulness of principal-agent (P-A) theory for the study of international relations through a set of well-integrated analyses of delegation to international (governmental) organizations (IOs). The editors begin with some useful, explicit definitions of key terms. They define delegation as a revocable “grant of authority” from one or more “principal(s)” to an “agent,” which enables “the latter to act on behalf of the former” in a specified domain and/or for a limited period of time. The agent's discretion in how to pursue the principal's objectives is a direct inverse function of the precision of the rules laid down by the principal. Agent autonomy, by contrast, is defined as the possible range of actions the agent can take contrary to the principal's interests, net whatever mechanism the principal may have put in place to control the agent. To the extent that an agent actually pursues his own interests contrary to the principal's, we see agency slack.
Ten substantive chapters then use this common terminology to adapt and extend P-A theory to analyze various specific cases of delegation to IOs, including the International Monetary Fund (J. Lawrence Broz and Michael Hawes, Erica R. Gould, Lisa L. Martin), multilateral development banks and aid agencies (Mona Lyne, Daniel Nielson, and Michael Tierney, Helen V. Milner), the European Union Commission and Parliament (Mark A. Pollack), the World Health Organization (Andrew P. Cortell and Susan Peterson), the United Nations Security Council (Alexander Thompson), international courts and the World Health Organization's Appellate Body (Karen Alter, Andrew P. Cortell, and Susan Peterson), and the Commission and Court under the European Convention on Human Rights (Darren Hawkins and Wade Jacoby).
One of the general insights from this volume is that P-A theory travels well from domestic to international delegation, across a wide range of issue areas. The same set of expected benefits motivate delegation, and cost or difficulty of monitoring increase agent autonomy, for instance in Gould's analysis of the IMF, just as for domestic agents.
So is there anything that makes international delegation distinctive? Some of the authors suggest that there is. Cortell and Peterson, for instance, argue that IOs-as-agents are more likely to have preferences that differ from those of the member states when the IO has a staff of its own rather than a staff seconded from national bureaucracies. David A. Lake and Mathew D. McCubbins, in their concluding chapter, suggest that the importance of information provision by third parties with the requisite expertise for keeping IOs-as-agents accountable creates incentives for states to retain specialized bureaucracies at the domestic level, even if most of their policymaking functions have been delegated internationally. There is no functional equivalent for this duplication in domestic delegation, which also suggests that international delegation may tend to be less efficient.
Another difference arises from Lyne, Nielson, and Tierney's distinction between an agent with multiple principals and an agent with a “collective principal,” consisting of a group of otherwise independent actors who jointly delegate authority to an agent. They argue that collective principals are particularly prevalent in international delegation, making international recontracting on average more costly. I believe they are right on this point, though decision-making rules need to be considered along with the structure of the principal to yield their conclusion: Unanimity requirements in international diplomacy or institutions (rarely, if ever, found in domestic political institutions) should make recontracting indeed on balance more costly for delegation to IOs. This in turn suggests that the ability of the principal to revoke the grant of authority might be better treated as a variable, rather than as a definitional dichotomy.
Delegation and IR Theory: As Gould and Pollack point out in their respective chapters, P-A models yield no point-specific predictions without being supplemented by assumptions about agent preferences derived from other theoretical traditions. This raises the question of how easily P-A models can be combined with or integrated into other theoretical traditions. One of the core assumptions of P-A theory is that the principal's decision to delegate, as well as the subsequent behavior of both principal and agent, are driven by the logic of instrumental rationality. This makes P-A compatible with rationalist theoretical traditions in IR, but hard to combine with approaches that reject the logic of instrumental rationality. Moreover, among the former, P-A theory seems to have a natural affinity with the liberal tradition in IR theory that distinguishes between the government of a state and sociopolitical groups at the domestic level and that takes the interests of the latter as the starting point of the analysis. The domestic electoral importance of U.S. money-center banks, for instance, explains both congressional support for increased funding to the IMF and IMF funding allocation in Broz and Hawes's chapter; and skepticism about overtly politicized bilateral aid by voters in donor countries (the presumed principals of elected politicians) explains the variations in the share of foreign aid allocated through multilateral institutions in Milner's chapter.
Other theoretical traditions in IR cannot recognize some of the theoretically most interesting benefits of delegation identified by P-A theory and might, therefore, have difficulty explaining the very existence of international delegation. Neorealism, for instance, which denies that international institutions have any effect independent of the distribution of power among states, must consequently also deny that states can use delegation to IOs to make more credible commitments or “lock in” policy—an important motivation for international delegation in the chapters by Pollack, Milner, and others. Neoliberalism acknowledges and in fact emphasizes these benefits of delegation, but by retaining the unitary-actor assumption still, cannot recognize blame avoidance or “shifting responsibility” as a motivation for delegation. In short, adopting a P-A perspective is not a neutral decision with respect to one's position in the broader theoretical debates in IR.
Organizations as Actors: In most of the chapters, change in a P-A relationship occurs only as a deliberate, conscious choice of the principals. Martin even explicitly argues that “IOs gain autonomy as a result of intentional state decisions to delegate [more] authority, not through a careless process driven by the staff” (p. 141). This seems at odds with one of the most important lessons of P-A theory for the study of international relations—emphasized in the introduction, the conclusion, and the chapter by Hawkins and Jacoby—namely, that international organizations are not just structures within which their member states interact, but that they can have agency independent of the member states. While Lake and McCubbins are right to note that governments may anticipate agency slack, organization theory as applied to IOs by Michael Barnett and Martha Finnemore, for instance, suggests that agents will seek to enlarge their portfolio or increase their autonomy. Hawkins and Jacoby show that such agents may try to increase their autonomy in various ways, including by reinterpreting the rules. And P-A theory's core assumption of incomplete information suggests that principals will anticipate such attempts at most imperfectly. Moreover, if it is correct that it is more difficult in international delegation for principals to rein in their agents, then such agent-induced changes may have greater persistence internationally than domestically.
A promising avenue for future research, therefore, will be to build on Hawkins and Jacoby's chapter and the work of Tim Büthe and Walter Mattli in order to develop theoretical models of delegation that allow for institutional change driven by agents, and to specify the conditions under which it can occur.
