Abstract
This article reviews the state of literature relevant to states’ burden-sharing behaviour within NATO. The purpose is two-fold: first, to delineate the different dependent variables and evaluate whether important questions have been left untreated, and, second, to assess strengths and weaknesses of the explanations that have been proffered. It is argued that while the system-level explanations capture major incentives to contribute, the domestic-level explanations are necessary in understanding specific decision-outcomes. The existing integrative models are superior to each explanation or level of analysis individually but tend to portrait domestic leaders as rather passive registers of international and domestic pressures. Empirically speaking, it is argued that more studies of contributions to distinct events, i.e. NATO operations, are needed, particularly focusing on cases of small states. The dependent variable of form of contributions is seemingly the least explored and may require incorporation of other theoretical arguments than utilized in existing works.
Introduction
Burden-sharing – ‘the distribution of costs and risks among members of a group in the process of accomplishing a common goal’ (Forster and Cimbala, 2005: 1) – has more often than not been a contentious issue within security alliances and coalitions. This has motivated a body of literature that explores the factors shaping states’ alliance contributions. In this article, the ambition is to review the state of literature relevant to burden-sharing within the North Atlantic Treaty Organization (NATO). 1 More specifically, the purpose is two-fold: first, to delineate the different dependent variables and evaluate whether important questions have been left untreated, and, second, to assess strengths and weaknesses of the explanations that have been proffered.
It seems a modest claim that the research field under review is just as important as ever. Burden-sharing debates have been an inherent part of NATO’s internal life ever since its inception, but have arguably become more heated in recent years. The debates over burden-sharing in Afghanistan and Libya are illustrative cases in point. Factors such as politically risky missions overseas, the lack of a common threat perception, and more generally a larger and less homogeneous group of member-states than during the Cold War have arguably added to the challenge of securing a ‘fair’ sharing of the alliance’s burdens. Overstretched armed forces and increasingly strained economies will probably add further to the burden-sharing challenge in the years to come. The research field should also be of interest with regards to development of wider theoretical generalizations, as it can add to our knowledge about the conditions under which states are likely to achieve successful security cooperation.
In order to avoid repetition of theoretical arguments, the main part of the article is structured according to the perspectives and hypotheses in what appears to be the main works on the topic under review. 2 The categorization largely follows that of Bennett et al. (1994), who in their study – treated below – incorporated most of the then existing theoretical arguments on states’ burden-sharing behaviour. The conclusions summarize theoretical and empirical strengths and weaknesses of the reviewed literature, and directions for further research are suggested.
It is argued that political leaders’ incentives to contribute are well captured by the systemic hypotheses. Specific decision-outcomes can hardly be fully understood, however, without taking the domestic explanatory variables into account. Existing integrative models are superior to each hypothesis or level of analysis individually, but a limitation inherent in these models is that they tend to portrait domestic leaders as rather passive registers of international and domestic pressures. Empirically speaking, it is argued that more studies of contributions to distinct events, i.e. NATO operations are needed, particularly involving cases of small states. The dependent variable of form of contributions is seemingly the least explored, and an increased understanding may require incorporation of other theoretical arguments than utilized in existing works.
The system-level perspective
This section explores the system-level explanations that are usually discussed and/or applied in studies of states’ burden-sharing behaviour.
Collective action
The majority of academic works on burden-sharing in the security realm has invoked the so-called collective action hypothesis. It is based on parts of Olson’s (1965) seminal work The Logic of Collective Action, and proposes ‘exploitation of the great by the small’ with respect to a public good. The underlying assumptions are that dominant states will benefit from the realization of the public good even if no other state contributes, and that small states expect dominant states to provide the good in adequate supply.
Most of the empirical studies invoking this hypothesis have been concerned with overall military expenditures in NATO. The early works indeed found the size of a member’s national income and the percentage of its national income spent on defence to be correlated (e.g. Olson and Zeckhauser, 1966). However, later studies demonstrated that the ‘exploitation of the great by the small’ within NATO was decreasing (e.g. Russett, 1970; Sandler, 1993: 447).
The hypothesis has also been included in a few works that seek to explain contributions to distinct events; for example in the above-mentioned 1994 article by Bennett, et al. Their study aimed to explain the size and form (military, diplomatic, or economic) of the contributions from the UK, Egypt, France, Germany, Japan, and the United States to the Operation Desert Storm coalition. As expected, the United States paid a disproportionate share of the total coalition costs. However, the hypothesis fared poorly in explaining all the other states’ behaviour; although they had an opportunity to free-ride on the US effort, they chose not to do so (Bennett et al., 1994: 72). Another example can be found in a 2004 article by Auerswald. He explored relative support (in terms of military and diplomatic contributions) for NATO’s 1999 Operation Allied Force by key members of the alliance: France, Germany, Italy, the UK, and the US. His findings were partly consistent with the expectations that the US would have been the most willing to contribute, and that the stronger European states would have been more willing to contribute than their weaker partners (Auerswald, 2004: 656).
These are only a few examples, but they illustrate that the hypothesis have been most successful in explaining why dominant states tend to accept the largest share of the burden. Overall it fails to explain outcomes other than disproportional burden-sharing and the relative static distribution of burdens (Ringsmose, 2009: 78; Thies, 2003: 7–8). Critics have in particular challenged the assumption that the output produced by alliances satisfies the definition of a public good – especially after the end of the Cold War; conventional defence is excludable and influenced by consumption (Ringsmose, 2009: 79). 3 Furthermore, the ‘100 percent common purpose’ assumption neglects the possible private benefits of contributing (Forster and Cimbala, 2005: 10–11); for example reducing the risk of exclusion from the benefits of collective defence.
Essentially, the hypothesis on collective action highlights incentives grounded in a quest for economic efficiency, but is rather apolitical and fails to account for other important motivations (Baltrusaitis, 2008: 52–53). Recognizing this weakness, many scholars have instead embraced a realist approach, emphasizing the role of geostrategic position on states’ propensity to contribute.
Balance-of-threat
The hypothesis on balance-of-threat suggests that the higher the threat perceived, the greater a state’s propensity to contribute will be. It is based on the neo-realist theory of Walt, as laid out in the 1987 The Origins of Alliances. He argued that states balance not against power, but rather against threats; a function of not only power, but also distance from potential aggressors, their (perceived) intentions, and their offensive capabilities (Walt, 1987: 5).
Bennett et al. (1994: 72) found that the balance-of-threat hypothesis helped explain the anomalous lack of free-riding in the Operation Desert Storm coalition. Auerswald (2004: 656), however, demonstrated that a focus on balancing against threats was not a good explanation for the observed behaviour in Operation Allied Force (Auerswald, 2004: 656). One could argue, however, that the Kosovo case perhaps was more of a ‘war of choice’ than later US/NATO-led operations. This point is elaborated upon below.
Furthermore, the hypothesis may be expected to have limited ability to explain the behaviour of minor allies. As argued by Ringsmose (2009: 77), the smallest among allies may sometimes have little or no incentive to balance external threats by internal means, as the link between increased military expenditures and improved security conditions is less clear than for larger powers. For the burden-sharing behaviour of small states in particular, alliance dependence may be a more valid explanation.
Alliance dependence
According to Snyder (1984: 466–471), states in alliances inherently face two fears. One is abandonment, whereby an ally de-aligns, realigns with the opponent, or simply fails to keep its promises. The other is entrapment, which means being dragged into a conflict central to an ally’s interests but not to one’s own. Reducing one of these risks tends to increase the other, creating an ‘alliance security dilemma’. Finding a solution to this dilemma requires making trade-offs between security and autonomy, and the choice that is made will primarily be determined by the relative dependence of the member states. The more dependent a state is, the more likely the costs and risks of abandonment will outweigh the costs and risks of entrapment (Snyder, 1984: 472). In a general form, the alliance dependence hypothesis suggests that the greater a state’s dependence on an ally, the more likely it is to accommodate its protector’s requests for support.
Strong support has been found for variants of this hypothesis in existing research. For example, Bennett et al. (1994: 72) documented that, even when the alternative explanations pointed against a contribution, dependent allies made sizable contributions. Ringsmose, moreover, examined three cases of political decisions in Denmark under the Cold War regarding the size of the armed forces, and demonstrated that Danish policy-makers were more concerned with their senior partners’ burden-sharing demands than with a threat emanating from the Soviet Union, and that the rise and fall of military spending clearly followed the pattern of allied pressures (Ringsmose, 2009: 91).
The domestic-level perspective
Existing studies indicate that system-level explanations fare pretty well in explaining political leaders’ incentives to contribute. However, they seem to struggle with explaining specific decision outcomes, for example the size and form of states’ contributions. To address this weakness, scholars have increasingly looked to the domestic sources of states’ burden-sharing behaviour (Kupchan, 1988: 326). There are few generally agreed upon hypotheses on the relationship between factors at the domestic level and states’ burden-sharing behaviour. Three unit-level factors can nevertheless be delineated.
State autonomy and public opinion
The expectations under this section, based on liberal theories, can be summarized in a broad way as follows: first, that states can be expected to be more prone to contribute when there is popular support for doing so; and, second, that institutionally weak executives should be more reluctant to use force.
Kupchan (1988) examined these factors while intending to explore the determinants of intra-alliance cooperation through a case study of NATO’s efforts to address security problems in the Persian Gulf in the beginning of the 1980s. He focused on political cooperation, measured in terms of mutual concessions and policy adjustments, and economic cooperation, primarily measured in terms of relative defence expenditures. He found that, whereas political outcomes were best explained by systemic theories, domestic factors were the most important influences on levels of defence spending. This suggested that decision-making on issues that have wider and more immediate implications for national constituencies (i.e. that involve extracting resources from society) are more subject to domestic pressures (Kupchan, 1988: 344–345).
Furthermore, the findings of Bennett et al. (1994) and Auerswald (2004) lend support to the expectations above. Chief executives that lacked public support were more reluctant to contribute militarily. Also, government autonomy implied a less hesitant posture toward the use of force. Especially interesting is the finding of Bennett et al. (1994: 73) that domestic leaders that were caught between dependence on the United States and a lack of public support opted for a financial rather than a military contribution.
There appears to be few other studies focusing on these factors in existing burden-sharing research, but those that are mentioned illustrate that the interests and power of domestic actors who are crucial for a government’s political survival should be included into the analysis if one wishes to understand whether burden-sharing demands are met. A government must take domestic political risks into consideration when deciding on whether or not to contribute, and domestic factors may also help to explain form of contributions.
The unit-level factors discussed thus far recognize that a government’s room for manoeuvre may be influenced by constraints internal to the state. Another perspective instead stresses that the state itself is not monolithic. Governments, according to the ‘bureaucratic politics’ model, consist of several actors with shared power, but only partly overlapping interests.
Bureaucratic politics
Bennett et al. (1994) explored a hypothesis on bureaucratic politics in their work on burden-sharing in the Persian Gulf War. Their theoretical discussion was built on the seminal 1969 article ‘Conceptual Models and the Cuban Missile Crisis’ by Allison. Three expectations were detailed and applied to their six country case studies: first, that government decisions are the outcome of bargaining among individuals and groups within the government; second, that players’ ‘stands’ on policy issues are, at least partially, shaped by their ‘seats’, or organizational roles; and, finally, that the outcomes of bureaucratic bargaining depend on top officials’ relative bureaucratic power (Bennett et al., 1994: 48).
The authors’ findings were in line with these expectations, but the process-tracing evidence, they recognized, was thin. In any case, one could ask if their findings truly indicate that ‘the pulling, hauling and bargaining below the President affect Presidential choice’ (Art, 1973: 474). In their study, other hypotheses explained the outcomes in a more parsimonious way. However, the broader literature (e.g. Halperin et al., 2006) on foreign policy-making attests to the fruitfulness of bureaucratic politics as a unit-level consideration. Indeed, in decisions on specific contributions, one could expect ‘what happens below the President’ to matter more. For example, in decisions about what force contribution(s) to offer an operation, it seems reasonable to assume that military officers often see a different ‘face of the issue’ from their political leaders, and that they have a bargaining advantage due to control over information that is needed to identify options (Allison and Zelikow, 1999).
Although ‘pulling and hauling’ may often be a mechanism at work, some common ideational ground can be expected to guide behaviour. This is recognized by the next theoretical approach.
Strategic culture
Why do states that share the same security environment and/or have similar domestic institutions diverge in their burden-sharing choices? Theories of strategic culture may be of help in explaining this empirical puzzle, as they explore how state interests are (socially) constructed (Katzenstein, 1996: 2). However, although numerous studies have been published on ideational explanations of national security policies, few scholars have systematically explored how theories of strategic culture may shed light on states’ alliance contributions. Nevertheless, a couple of recent works on NATO members and the use of force may provide some hints on the fruitfulness of this approach.
The first work is a 2005 article by Rasmussen. He asked why Denmark became more ready to project military force following the end of the Cold War, and convincingly argued that Denmark came to regard the use of armed force in new ways following the end of the Cold War (Rasmussen, 2005: 67).
In the second work, a 2008 article by Noetzel and Schreer, the challenges of the Bundeswehr thus far in Afghanistan were explored. They argued that domestic strategic culture was an important explanatory factor for why the German government had failed to meet its NATO allies’ expectations thus far. They showed that, as a result of successive military deployments during the 1990s, the general issue of deploying the Bundeswehr was no longer a source of contention in German domestic politics. On the other hand, however, they demonstrated that the actual use of force in combat was still subject to the German ‘culture of restraint’; elites remained sceptical about the utility of force in reaching political goals (Noetzel and Schreer, 2008: 219).
This is not the place to try settling disputes on, for example, how strategic culture should be conceptualized, how one should go about studying it, or whether it should be treated as a cause or context of state behaviour. In this article, it suffices to note that strategic culture may allow one to say something about a country’s willingness to use of force, and for what purposes.
Integrative models
The first attempt to integrate systemic and domestic factors and explain contributions to an international operation was seemingly performed by Bennett et al. (1994). First, they compared the explanatory power of the three externally driven hypotheses and two of the internally driven hypotheses treated above. They found that external pressures explained political leaders’ incentives to contribute to the Operation Desert Storm coalition, but that internal constraints accounted better for their ability to contribute and the form contributions took. Second, after demonstrating that no single hypothesis could explain every state’s burden-sharing behaviour, they proffered a security decision-making model that showed how their hypotheses could be combined into a single explanatory framework. 4 Above all, their work can be characterized as an important attempt at ‘bridge-building’ between competing schools of thought, as it incorporated the main existing theoretical arguments about states’ burden-sharing behaviour.
The second existing integrative model for explaining states’ burden-sharing behaviour and expand burden-sharing literature to individual operations can be found in the 2004 article by Auerswald. He developed a model that incorporates core concepts of theories of collective action, balance-of-threat neo-realism, public opinion, and government institutional structures, presented in a staged, conditional manner. Auerswald found that the integrated model did a better job of explaining states’ behaviour in Kosovo than each theory individually, and that it did so in a rather parsimonious way. However, although Auerswald’s model provided a parsimonious explanation for burden-sharing among a few countries in the 1999 Operation Allied Force, its generalizability may be limited. First, the concept of threat is omitted from his integrated decision-making model, based on an assumption that no group member’s survival is threatened (Baltrusaitis, 2008: 108). For the Kosovo case, this may have been a valid assumption. However, as illustrated by the actions of the US in the aftermath of 11 September 2001, not every military intervention is perceived by state leaders to be a ‘war of choice’. Second, and on a more general note, the model leaves little room for different state motivations for contributing.
A limitation inherent in these models of states’ burden-sharing behaviour should be mentioned. They tend to portrait domestic leaders as rather passive registers of international and domestic pressures, but one could argue that their role as ‘gatekeepers’ 5 between the international and the domestic arena provides the opportunity to try influence outcomes. As stated by Thies (2003: 5–8, 162), NATO members can usefully be thought of as burden-shifters intent on manipulating alliance relationships to maximize gains at both the international and the domestic level.
It may be that exploring states’ burden-sharing behaviour within the framework of a ‘two-level game’ is fruitful. The term serves as a metaphor for the entanglement of domestic politics and international relations (Putnam, 1988). Although not a theory as such, the added value of Putnam’s approach compared with the works mentioned above is exactly the recognition of the potential for purposeful behaviour on the part of domestic leaders.
An important concept in the two-level game framework is that of a domestic ‘win set’: the set of potential agreements that would gain the necessary majority among the constituents when simply voted up or down (Putnam, 1988: 437). In other words, the win set sets the outer limits to governmental choice. Bosold and Oppermann (2006) analysed UK and German governmental discourses on the invasion of Iraq in 2003. More specifically, their paper concerned the question of whether or not to be part of the US led coalition, and how the German and UK governments tried to mobilize sufficient domestic support for their preferred decision. Drawing heavily on the above- mentioned work by Putnam, the essence of their theoretical argument is that win sets are not to be conceptualized as static restrictions on government choice; governments can attempt to shape their win sets through purposeful behaviour. In their own research, they focused on the use of language. First, they identified the crucial domestic veto-players in Germany and the UK respectively. ‘Veto-players’ can be defined as ‘actors whose agreement is required for a change of the status quo’ (Tsebelis, 2002: 17). Second, they analysed the German and UK governmental discourses addressed at these veto-players. Bosold and Oppermann found that the German government’s refusal to go to war with Iraq and the UK decision to be part of the US led coalition were accompanied by diametrically opposed domestic discourses directed at fundamentally different audiences; this can be interpreted as efforts to influence win sets, i.e. build sufficient domestic support to not risk losing office.
Of course, the use of language is but one example of a tactic that governments can employ to try influence the size of win sets. Furthermore, flexibility at home may not be the only rationale; the size of win sets also affects the bargaining power of governments on the international level. A bargaining perspective based on the concept of win sets can possibly be a fruitful one in studies of NATO burden-sharing.
Conclusions
Based on existing studies, this article has argued that the traditional approach to studying states’ burden-sharing behaviour – the collective action hypothesis – fails to explain outcomes other than disproportional burden-sharing and the relative, static distribution of burdens. To account for other important motivations, the realist concepts of balance-of-threat and alliance dependence have proven useful. Seen in totality, however, the systemic hypotheses, which dominates the existing burden-sharing literature, struggle with explaining specific decision outcomes; their strength lies in explaining political leaders’ incentives to contribute.
Particularly with regard to questions concerning other aspects of states’ burden-sharing behaviour than broad tendencies such as propensity to contribute, there seems to be much to gain by taking domestic explanatory variables into account. Decisions on the size and form of contributions are too specific to be adequately addressed by a grand theory of international politics.
Integrated models are an alternative to theories that direct one’s attention to either systemic or state-level factors influencing a state’s burden-sharing decisions. They are hardly parsimonious, but arguably better at capturing the complexity of real-world decision-making. Existing integrative models for explaining states’ burden-sharing behaviour do a good job of incorporating a variety of theoretical perspectives but tend to portray domestic leaders as rather passive registers of international and domestic pressures.
Empirically speaking, a large majority of the existing works are concerned with overall military expenditures. This is hardly surprising, given that the core focus in the burden-sharing debates during the Cold War was who was bearing their fair share of the burden of contributing to NATO’s common defence against the Soviet Union. The primary indicator for assessing efforts reflected in these debates was states’ defence spending as proportion of their gross domestic product (Chalmers, 2001: 570; Ringsmose, 2007: 19). However, changes in the international system following the collapse of the Soviet Union have influenced perceptions among NATO members of the criteria for burden-sharing. After the end of the Cold War, and especially after the terrorist attacks against New York and Washington on 11 September 2001, international military operations has moved front and centre in transatlantic security policy. The number of operations has increased, the goals and objectives have become more ambitious, the rules of engagement more robust, and the missions more politically risky given the increased probability of events such as casualties and mission failure. In the post-Cold War context of ‘collective security’, forces for overseas operations are on top of the hierarchy of possible contributions (Forster and Cimbala, 2005). This shift in the transatlantic burden-sharing debate has awakened an interest in explaining contributions to distinct events, but thus far only a handful of works have been published.
With regard to the works that do focus on contributions to distinct events, two points are especially worth mentioning. First, more country studies within a wider operational context are needed to validate the plethora of hypotheses discussed above. In particular, more empirical studies involving small states are needed. More circumscribed strategic interests and limited resources than those of great powers may yield different decision-making dynamics. Analysing in more detail the small states’ incentives, disincentives, possibilities, and constraints for international participation is therefore clearly of importance in understanding the full range of burden-sharing outcomes.
Second, the form of contributions is particularly understudied. True, scholars have begun to uncover the conditions under which a state is likely to provide military, economic, and/or diplomatic support. It can be argued, however, that one of the more important questions in today’s context has been left largely untreated: what explains states’ decisions on what forces to contribute? For example, the NATO-led operation in Afghanistan (ISAF) has clearly demonstrated the importance of examining not only the number of troops deployed, but also what forces are contributed and what tasks they are able to perform. Countries that have deployed combat-capable forces and have been willing to accept the more risky missions have clearly enjoyed a higher degree of receptiveness with allies than what can be explained in terms of size of troops only.
A discussion of theoretical arguments applicable to states’ decisions on what forces to contribute is beyond the scope of this article. Some comments are nevertheless in place. A potentially relevant factor not covered explicitly by any of the theoretical approaches discussed above but recognized by the ‘organizational processes’ approach is the ‘fixed programs’ that limit the range of options that leaders can choose from (Allison and Zelikow, 1999: 164). Furthermore, even when options for contributing forces exist, most countries face difficult trade-offs in responding to demanding calls made on their armed forces for overseas operation. To understand how such trade-offs are solved, more extensive studies into domestic bureaucratic processes – particularly involving military organizations – are needed. This may also shed light on questions of acute interest in today’s security context, such as why NATO-members time and again fail to deliver their collectively agreed-upon force levels.
Footnotes
Acknowledgements
The author is grateful to Magnus Petersson and Jan Ångström for their valuable comments on earlier drafts of this article.
Funding
This article received no specific grant but the underlying research benefited from a PhD scholarship provided by the Norwegian Ministry of Defence.