Alliances

Carlo Masala
Alliances have been an important element of security policy since long before the rise of
the modem nation-state. Tribes, princedoms and later nation-states have always sought
out allies with whom they could pool their resources in the pursuit of cotrunon goals.
Therefore, the history of alliances is as long as the history of relations between cohesive
units of human coexistence.
In this chapter, an alliance is regarded - in accordance with the definition of Stephen
Walt (1987: 1) - as a 'fonnal or infonnal relationship of security cooperation between at
least two sovereign states' . Therefore, alliances can take the shape of either fonnal or
infonnal intemational institutions (Duffield 2007). Furthennore, their structure can correspond
to one of two ideal types (Weber 1 997: 33) - hierarchical or egalitarian. 2 The
fonner type is characterized by significant imbalances of capabilities between alliance
members, whilst in the latter type, power is distributed more or less evenly among most
members.
Hierarchically structured alliances can be further differentiated into hegemonic and
imperial alliances. This distinction refers to the way the strongest power in the alliance
exerts its leadership. If the strong state leads with the consent of the smaller powers, their
relationship is considered a hegemonic one (c£ Triepel 1 938) . Arrangements under
which the relationship between the strongest state and the other alliance members is based
on coercion (as was the case, e.g. in the Warsaw Pact) are considered imperial alliances.
Although alliances are a core element in the history of Intemational Relations and the
concept itself is a 'key tenn' (Modelski 1 962: 773) in academic discourse, alliances are at
the same time 'understudied' (Snyder 1997: 1 ) . There are plenty of studies on alliance
management, but research is lacking on why alliances are fonned and when they dissolve,
which is astonishing given the fact that 'alliances are apparently a universal component
of relations between political units, irrespective of time and place' (Holsti et al.
1973: 2) .
One of the reasons for this puzzle is that the specific issue of alliances is difficult to separate
from other fields in the discipline of Intemational Relations. It is impossible to discuss
alliances without referring to more general theories of Intemational Relations, e.g. realism
(Morgenthau 1 948), neorealism (Waltz 1 979), neoliberal institutionalism (Keohane 1 984)
or constmct1v1sm (W endt 1 999), or touching upon so-called 'substantive issues' (Carlsnaes
et al. 2002: iv) , e.g. conflict theory (Zartman 1 985) or deterrence (Freedman 2004).
Because the topic of alliances is inextricably intertwined with other fields of the discipline,
theorists have either focused on a particular alliance (mostly on NATO) or
developed partial theories focusing on particular aspects of alliances.
So far, there have been only three attempts to create anything close to a comprehensive
theory of alliances. The first attempt was George Liska's (1 962) 'Nations in Alliance',
followed 25 years later by the seminal work of Step hen W alt (1987) on 'The Origins of
Alliances' and, another decade later, Glenn Snyder's (1 997) work on 'Alliance Politics'.
All three studies have their limitations. Liska's attempt to elaborate a theory of alliances,
although it gives some useful insights on the creation of alliances and patterns of cooperation
among its members, clearly suffers from its anecdotal character. Walt's work is
limited, as the title indicates, to the origins of alliances, and Snyder's book on alliances in
multipolar systems focuses on the management of member relations.
The lack of a discrete body of knowledge about alliances became very much apparent
during the 1 990s, when the question of NATO's survival after the end of the Cold War
led to an academic battle between various schools of thought3 without any tangible
results in tenns of cumulative knowledge (Lakatos and Musgrave 1970) or paradigm shift
(Kuhn 1962) . Much of the debate seemed to be self-referential and focused on the
affinnation of the different research progra1mnes, rather than striving for the enhancement
of our knowledge about alliances.
This chapter tries to cut through the current confusion about alliances by providing an
overview of different, occasionally competing, explanations. The following addresses
three issues that are fundamental for our understanding of alliance politics. Firstly, why do
alliances exist? Secondly, how do they function? And thirdly, when do alliances dissolve, and
under which conditions do they survive?
The origins of alliances
There is a widespread agreement among the neorealist, neoliberal institutionalist and
constmctivist schools of thought that states do act and interact in the absence of a centralized
authority capable of providing protection and the 'redress of grievances' (Grieco
2002: 65) . Each of these approaches to understanding alliance fonnation is discussed in
more detail below.
Balancing and bandwagoning
For neorealists, alliances are tools for balancing where states are unable to establish equilibrium
by relying on their own means. 4 Therefore, states use alliances as an instrument to
maintain or improve their relative power position globally or regionally. Neorealists
believe that states decide to fonn or to j oin alliances based on exogenous, not endogenous
motivations, because ' [a]lliances are against, and only derivatively for, someone or
something' (Liska 1962: 12) . Alliances therefore can be regarded as a particular outcome of a
conflict. By building alliances, states try to maximize their capabilities to counterbalance
the overwhelming power of another individual state or group of states.
Stephen Walt has modified this neorealist account of alliance fonnation somewhat. By
including perceptual and behavioural variables in the body of neorealist theory, he argues
that states do not aim to counterbalance power per se, but the power of actors they
perceive as threatening (Walt 1987: 21-28) . However, seeking to establish equilibrium is
only one option available to states. If states that feel threatened are unable to pursue a
balance-of-power strategy on the strength of their own capabilities and have no potential
allies, they may be forced to 'accommodate the most imminent threat' (Walt 1987: 30)
to their security. From this point of view, bandwagoning is a strategy to avoid becoming
the victim of a threatening state or to enj oy the anticipated spoils of victory. But, in
general, Walt (1987: 33) concludes, states are inclined to balance rather than to bandwagon,
since bandwagoning always involves an unequal exchange where one state (the
weaker one) accepts a subordinate role.
While W alt considers bandwagoning to be a kind of anomaly within the neorealist
research framework on alliances, Randall Schweller (1 994) regards it as being compatible
with realist assumptions about state behaviour, if the motivation of a state that joins the
stronger rather than the weaker side is revisionist, meaning that the state in question is primarily
concerned with destroying the current order and securing additional gains. 'Many,
therefore, choose to bandwagon with revisionist great powers bent on constructing a
new international system; they are "power-maximizing states"' (Zongyou 2006: 1 96).
From a neorealist perspective, alliances are a fonn of 'regression' in conflict regulation
behaviour (Singer 1 949). Such regression (which refers to an elimination of tensions
through reduction of complexity in the relations between alliance members and the
threatening power[s]) among conflict parties goes hand in hand with integration, or
strengthening of the alliance's overall fabric as a result of intensified relations among its
members. Therefore, members of an alliance face an additional constraining effect on their
action and interaction, because, from a neorealist perspective, an alliance does not abolish
the constraining effects emanating from the anarchical structure of the international
system on state behaviour, but merely modifies them.
Cooperation as reward
While neoliberal institutionalists do not deny that states are acting and interacting under
conditions of system-wide anarchy, they do not attribute the same effects of anarchy to state
behaviour that realists/neorealist do (Masala 2005: 92; Keohane 1984). Anarchy, therefore, is
not an obstacle to lasting cooperation among states. States engage in alliances because, as
selfinterested
actors, they anticipate a mutually rewarding exchange among the members of
an alliance (Stein 1 990: 7) . But what exactly are those rewards that states expect to reap?
Firstly, states create alliances in the expectation that the alliance members can achieve a
certain degree of cooperation. As long as the costs for the creation of alliances do not
outweigh the perceived benefits from cooperation, states are eager to cooperate (W allander
2000: 706) . From this point of view, an external threat can trigger alliance formation,
but there are further advantages beyond the mere engagement in a counterbalancing
effort that make alliance membership an attractive proposition for states. From a neoliberal
institutionalist viewpoint, alliances provide reciprocity, make members accountable
for their actions, and contribute to the creation and maintenance of cooperative
security strategies. They also reduce uncertainty by providing credible infonnation on the
behaviour of member states and make state behaviour among members of the alliance
more predictable by developing nonns and rules that regulate it (Wallander et al. 1999:
3£). Although there is a price to pay for j oining such institutions, as they impose constraints
upon state strategies and have an influence on state preferences, it is a price that
states may consider cmrunensurate with the expected benefits. In fact, it is precisely
because institutions and actions undertaken by them are costly (politically and economically)
that they are credible and therefore can be valuable to self-interested states (Wallander
and Keohane 1999: 30). To summarize: from an institutionalist perspective,
alliances offer their member states many advantages, which guarantee that alliances persist
beyond the conditions in which they were created.
The domestic factor
A third and relatively new account ofhow and why alliances are created is offered by the
liberal school. Liberalism in general 'seeks the roots and causes of external behaviour in
domestic structure and process' (Muller 2002: 376) . In general, liberals share the belief
that states engage in alliances when there is a convergence of national preferences created
by domestic coalitions (Risse-Kappen 1 9 9 1 ; Moravcsik 1 998) . Risse-Kappen believes
that the creation of NATO was not dependent on a real or perceived threat, or even a
constructed one. However, Risse-Kappen and Moravcsik both focus on democracies
only, leaving aside the possibility of preference conversion amongst non-democratic
states. Thus, liberals are able to explain why NATO was created, but have difficulties
explaining the foundation of the Warsaw Pact Organization, which was fonned under
Soviet leadership amongst Socialist or Communist countries.
Common identity and ideas, values and norms
So far, not much work has been done by constmctivists with regard to alliances. The most
elaborate and sophisticated constructivist-inspired analysis on the question of why alliances
(specifically NATO) are created is offered by Thomas Risse-Kappen (1 996) .
According to Risse-Kappen, NATO was not created as an effort to counterbalance the
Soviet material threat. Rather, NATO represents the institutionalized fonn of cmrunon ideas
and worldviews about the coming intemational order after the Second World War shared by
the founding states of the alliance (ibid. : 387). Those with similar worldviews and fundamental
beliefS founded an alliance whose guiding principles were consistent with their values
and nonns. In this respect, NATO is an alliance of identity that is not threat-based, but
reflects a relationship between states based on a cmrunon understanding of their shared traits.
With his emphasis on shared beliefS, Risse-Kappen paved the way for others to apply the
same concept to other alliances. Michael Bamett (2002) in his article on alliance fonnation in
the Middle East challenges directly Stephen Wait's assertion that ideologies have played an
'important but ultimately limited role' (Walt 1 987: 203). Bamett argues instead that identity,
and thus ideology, is a key element to social and political interactions, that has to be examined
closely to understand why states fonn an alliance. 'A starting point for the study of
identity is a belief that social groups need to establish a positively valued distinctness from
other groups in order to provide members with a positive identity' (McCalla 1 996) .

How do alliances work? Alliance management

While our knowledge on why and how alliances are created is limited, the literature on
the management of relations among alliance members is manifold. There are a number
of partial theories focusing on different aspects of alliance management.
The distinction between different types of internal alliance structures is important for
the analysis of alliance management. Depending on their internal structure, alliances may
face different management challenges. This section touches on three central aspects of
alliance management: (a) hegemonic stability; (b) the alliance internal security dilemma;
and (c) alliance cohesion. While the last two aspects are important for understanding
alliance management within hegemonic/imperial as well as egalitarian alliances, the first
occurs only in hegemonic alliances.
Hegemonic stability
The internal structure of an alliance has implications for the management of intra-alliance
relations. In a hegemonic alliance, smaller states are only willing to subordinate themselves
to the leadership of a bigger state, thereby constraining their sovereignty, if the
hegemon provides a public good that the smaller states are unable to produce sufficiently
by themselves. This is the core of the hegemonic stability theory (Kindleberger 1 986).
There are, however, differences in explaining why hegemons provide public goods.
Neorealists emphasize that it is in the interest ofhegemonial powers to create alliances, to
shape them and to exert their influence upon other states through fonns of institutionalized
cooperation ( Gilpin 1 9 8 1 ) . N eo liberal institutionalists on the other hand, emphasize
that the hegemon's behaviour is driven by enlightened self-interest. The hegemonial
power provides the public good and lets others participate because it is in the interest of
all member states of an alliance (Keohane 1 984) . The difference between the neorealist
and the neoliberal institutionalist reasoning is that neorealists would always highlight the
fact that the hegemon gains a political advantage from the provision of a public good for
free, and only provides this good for as long as it is in the interest of the leading power
and it is able to use the other alliance members for its own purposes. From a neorealist
perspective, public goods are provided because the hegemon is interested in relative
gains. Neoliberals instead would argue that the provision of a public good by a hegemon
is not linked to the political exploitation of other alliance members. From a neoliberal
perspective, the hegemon is interested in absolute gains and will therefore refrain from
exploiting the fact that its relations with the other alliance members are asytrunetric in
nature.
Both schools of thought would agree, however, that the type of relationship between
the hegemon and its followers as outlined above entails the risk of free riding. If the
hegemon provides a collective good, smaller states do not see the need to increase their
efforts to contribute to the production of this public good. In the long tenn, this creates
tension within the alliances. The difference between both schools of thought is, however,
that realists argue that tensions only occur in the absence of a cmrunonly faced
threat, while neoliberal institutionalists see tensions as a pennanent feature of relations
between alliance members.
The internal security dilemma
A feature that is common to both the hegemonic and the egalitarian alliance is what
Glenn Snyder - using Robert Jervis's (1 976: 63) concept - has described as the alliance
internal security diletruna (Snyder 1997: 180-83) . In an alliance, as in the international
system in general, the absence of a supranational authority leads to a situation where
many of the steps pursued by states to bolster their security have the effect - often
unintended - of making other states less secure. Since alliance members can never be
certain of other alliance members' future or present intentions, they embark on policies
vis-a-vis their alliance partners aimed at enhancing security.
Snyder's starting assumption is that the interests of alliance members never fully converge.
Since alliance members have cmrunitted themselves in a more or less binding way
to certain goals, every alliance member faces two potential dilemmas. The first is the risk
of entrapment (Snyder 1 997: 1 98 1 ) . Entrapment refers to a situation where an alliance
member (A) faces the choice of supporting another alliance member (B) as a result of
treaty obligations, although A has no particular interest in supporting B, or staying out of
a conflict. Country A fears involvement in a conflict that does not involve its vital
interests. However, if A, despite its commitments, stays out of the conflict, it may risk
defection by its ally.
The second risk Snyder describes is the fear of abandonment. Abandonment characterizes
a situation from the perception of state A, which has a particular interest in a conflict
with a non-alliance member, but cannot be sure of the active support of other allies.
To attenuate both diletrunas by avoiding entrapment as well as abandonment is one of
the biggest tasks in the management of an alliance. Both the management of the alliance
internal security diletruna as well as the structural problems inherent to hegemonic stability
theory point to a much larger problem that alliance members face in the management of
their relations, namely how to ensure alliance cohesion.
Alliance cohesion
From the neorealist point of view, alliance cohesion, whether in a hegemonic or an
egalitarian alliance, depends on the degree of the external challenge. The bigger the
threat alliance members face (or perceive) , the greater alliance cohesion will be (Mearsheimer
1 990) . There is disagreement, however, as to the role that the hegemon plays in
maintaining alliance cohesion. While some authors regard it as crucial (Gowa 1 999),
others are sceptical, particularly concerning the phase when the cmrunonly perceived
threat decreases (Masala 2003). When the threat decreases, smaller alliance members are
less and less willing to subordinate themselves to a hegemon. The simple causality of
neorealist reasoning is as follows: high levels of threat perception lead to high alliance
cohesion, while conversely, low threat perception levels result in low alliance cohesion.
From a (liberal) institutional perspective, alliance cohesion is guaranteed by the multiple
advantages that alliances provide for their member states. The reduction of transaction
costs; the ability to control other alliance members; the access to infonnation about
the intentions and behaviour of alliance partners; and the iterated games that are played
within the cooperative framework of an alliance - all of these advantages outweigh the
potential costs and frictions emanating from different national preferences (Keohane and
Nye 1 993) .
Constructivists (similar to neoliberal institutionalists) argue that alliances have cohesion
because of the 'republican liberalism linking domestic polities systematically to the foreign
policy of states' (Risse-Kappen 1996: 358) . Among liberal democracies, coherence is
guaranteed because liberal democracies are more inclined to cooperate closely with other
liberal democracies. This is because liberal democracies do not regard other liberal
democracies as a potential threat. Among liberal democracies, there is no fear that
cooperation will be exploited by allies in the future; therefore, the problem of relative
gain distribution (Grieco 1 990) does not exist. Additionally, constructivists point to the
fact that democracies are characterized by transparency; high audience costs; consistent
policy behaviour; civilian control of the anned forces (which is important for military
alliances between democracies) ; and the capacity to make enduring commitments (Lai
and Reiter 2000) - all of which are factors that facilitate closer cooperation between such
regimes. To summarize: the more similar the regimes are that cooperate, and the more
their values converge, the more cohesive the alliance they fonn will be.
As with the question of why alliances are fonned, there is no clear-cut answer to the
question of how they are managed.

Why do alliances end, why do they survive and how?

The issue of how alliances survive and why they end was hotly debated among scholars
of International Relations during the first half of the 1990s. It should not come as a
surprise that the results were, at best, mixed.
The linchpin for this debate was, of course, the question of whether NATO would
survive the end of the Cold War. Unsurprisingly, neorealists were very sceptical concerning
the future of alliances after the disappearance of the cohesive force that holds
them together; namely, the cmrunonly perceived threat. Scholars like Waltz and Mearsheimer
were quite outspoken about the future of alliances, especially of NATO, after
the demise of the Warsaw Pact and particularly after the collapse of the Soviet Union.
They believed NATO had become an anachronism (Waltz 1 990: 2 1 , 2000: 1 8 ; Mearsheimer
1 990) . Waltz believed that the reason why NATO still existed, even ten years
after the fall of the Berlin Wall, was the fact that the US had an interest in maintaining its
'grip' (Waltz 2000: 1 9) on developments in Europe. Stephen Walt j oined his fonner Ph.D.
supervisor in this rationale by pointing out that US hegemony could well explain why
NATO remained one of the main relevant political institutions in transatlantic affairs
(Walt 1 997: 1 7 1 ) . But even the continuation of US hegemonic policy, according to
Waltz, would not prevent NATO from becoming an irrelevant institution, because other
alliance members would be less and less willing to accept US supremacy within the
institution.
On the other side, neoliberal institutionalists have been extremely optimistic with
regard to the possibility of cooperation within alliances even without an external threat
or a hegemon (Haftendorn et al. 1 999) . Besides the already mentioned positive effects
that alliance members enj oy by virtue of their alliance membership, institutional inertia
contributes to the prolongation of cooperation within an alliance.5 Furthennore, as long
as alliance members have an interest in keeping their cooperation alive (Keohane 1984:
31), the chance that alliances may survive and even adapt to a new environment - even
in the absence of an overwhelming threat - are quite high.
An argument as to why NATO survived the end of the East-West conflict that is fully
in line with the basic tenets of neoliberal institutionalism has been developed by Celleste
Wallander (2000). Institutions with general assets,6 she argues, will be adaptable to new
problems. Because the assets are not specific to a given relationship, location or purpose,
using them for new purposes will be low-cost and broadly effective (Wallander 2000: 709).
NATO, she concludes in her empirical analysis, has been successful in adapting its assets
to a new security environment (ibid: 732) that is not characterized by a unifying threat.
According to the constructivist point of view, the survival of an alliance depends on
the continuation of the underlying reason for the alliance's existence, which, as discussed
above, is not a cotrunon threat, but the perception of having a common destiny.
Therefore, alliances can survive major changes in their environment if their members still
feel that they belong together and share the same nonns and values. From a constructivist
perspective, the end of the East-West conflict does not mark the end of NATO, since it
did not tenninate the cmrununity of values. Frank Schitrunelfennig (1999) argued that
after the fall of the Berlin Wall, the Western value community extended into the newly
democratic countries of Eastern Europe, and NATO played (and still plays, according to
this view) a crucial role in socializing these states. As long as alliance members build a
security cmrununity (Deutsch et al. 1 957f and have a sense of belonging to it, security
cooperation will continue.

Conclusion and outlook

This chapter has attempted to remove some of the confusion surrounding the study of
alliances by systematically structuring the existing literature into various categories along
three important questions.
We are left with the insight that, although alliances are important in structuring relations
between states, we still do not know much about them. The lack of knowledge
about alliances also has policy implications. One important question concerns the purpose
of alliances (from an alliance member perspective). Do they serve as an instrument
for (counter-) balancing real or perceived threats and risks? Are they instruments for the
management of relations between member states? Or are the ties that bind members of
an alliance together based on the notion that the states in question belong to a community
of a shared identity? The answer is that each of the competing perspectives
sketched above captures important aspects of alliances, and it would be 1nisleading if
policymakers confined their thinking about the role of alliances to only one perspective.
Heads of state and diplomats should be aware of the role that neorealists allocate to
power, but also take into account the domestic and institutional aspects as well as the
ideational foundations that scholars from the neoliberal institutionalist, liberal and constructivist
camps assign to alliances. A syncretistic conception of alliances tnight result in a
better understanding of the 'real-world developments' of alliances than the one resulting
from a continuous battle for supremacy between various schools of thought.
Secondly, all three maj or schools of IR theory focus on the question of whether or
not alliances ultimately wither away. As far as this issue is concerned, neorealism, neoliberal
institutionalism and constructivism offer relatively static explanations.
What is lacking is a dynatnic approach focusing not on the question of why NATO
still exists, but on how NATO is developing. All three of the approaches outlined above
shed some light on important issues, but none manages to explain comprehensively the
developments and dynatnics that have taken place, especially in NATO since 1 990. The
central hypothesis of a theory focusing on change in alliances is borrowed from neofunctionalist
theory, according to which the fonn of an institution follows its function
(Mitrany 1 976). There is a direct relation between the external environment and the degree
of institutionalization within an alliance. If substantial changes in the environment are
taking place, a high degree of institutionalization within an institution may give rise to
conflicts among its members (Simmel [1 9 1 8] 1 984: 38) if the commonly perceived threat
has withered away. If member states of an alliance have an interest in keeping the alliance
alive, institutional engineering is required to adapt the alliance to its new environment.
In this process of institutional re-adaptation, the internal security diletruna poses a major
challenge to alliance members. At the theoretical level, there are two ways of attenuating
this diletruna: either through a higher degree of institutionalization or through a weakening
of institutional ties among member states. The first option, which entails the
strengthening of commitments among alliance member states, can be successful where
there is an increase in threat perception. If the new alliance environment is characterized
by a cmrunonly perceived threat, an integrative strategy (reinforcing institutional ties) might
be an appropriate way to ensure that fears of entrapment or abandonment are assuaged.
The second strategy seems to be adequate if member states cooperate in an environment
that is characterized by 'risk-diffusion' (Masala 2003: 1 3) and an interest on the part
of member states in maintaining institutionalized cooperative relations. In such a situation,
a lessening of institutional constraints might give member states a higher degree of
freedom of action while at the same time minimizing any fears of abandonment or
entrapment that might exist.
Such a flexible alliance (Masala 2003: 32-36) could - from a member-state perspective constitute
an appropriate response to the changed environment and thereby guarantee
the continuation of institutionalized security cooperation in the fonn of alliances even in
the future. This approach to studying changes in alliances has two distinct advantages:
Firstly, it borrows important insights from most of the theories dealing with alliances
(or at least some important aspects of them) and tries to combine them into a line of
inquiry that is not only relevant from a political-science point of view, but also from a
political point of view.
Secondly, it leaves open the question of whether an alliance will survive or not. Both
traj ectories are possible, and the answer depends on real-world developments within
alliances and among their members.