The paradox of anarchy

Transcription

1 Department of Political Science The paradox of anarchy Why anarchy is a rational choice Martin Lundqvist Independent Research Project in Political Science, 30 credits Master s Programme in Political Science Year, Term: 2016, Fall Term Supervisor: Magnus Lundgren Word count (excluding appendices): 19280

2 Abstract A central paradox in neorealism is that the absence of world government is assumed to imply a dangerous security dilemma, and yet few realists have argued in favor of world government while great powers have historically resisted delegation of military force to supranational institutions. If international anarchy causes costly security competition and war, powerful states should have a strong incentive to unify and neorealist explanations for why states resist peaceful unification are either underdeveloped or implicit. This paper develops a rational choice realist model which suggests that anarchy is not a structural constraint which forces states to compete intensely for power, but a rational choice that state leaders make to avoid the costs of world government. The model assumes that state leaders face a delegation constraint which implies a tradeoff between eliminating the military burden necessary to deter attacks in anarchy, and abstain from world government to avoid forced redistribution of material resources pushed by poorer states or risk that the world government might turn tyrannical and coerce the subjects it has been mandated to protect. The paper uses deductive method to deduce actor preferences and illustrate the plausibility of the central predictions of the model. The model predicts that income equality, democratization, nationalism and military defensive advantage makes anarchy stable, a condition under which states have little reason to unify. On the other hand, high income inequality, lack of democratic accountability, territorial revisionism, and military offensive advantage make anarchy costly, but unification more difficult to achieve given the underlying conflicting preferences. Hence, states rather take the risk of fighting in hopes of eliminating future military competition than to agree on a world government where redistribution and cultural conflicts are likely to make wealthy great powers with modest population sizes worse off than they would be in anarchy. The paper concludes that anarchy is not a tragedy that makes world government impossible, but world government is a potential tragedy that makes anarchy rational.

3 Contents Chapter Introduction and research puzzle Research questions and strategy... 2 Chapter Previous research: introduction The virtues of anarchy Territorial competition and the cost of anarchy Hobbesian contract between men but not between sovereigns International cooperation and institutional design The historical evolution of international anarchy Chapter Method and operationalization: introduction The congruence method Deducing actor preferences Delimitations, reliability and validity Chapter Theoretical framework: Introduction Fundamental assumptions Theoretical model Power balancing and the cost of anarchy The role of uncertainty Chapter Analysis: introduction Taming the leviathan Military technology and the offense/defense balance Territorial greed and the cost of anarchy relative to world government Income inequality and the redistribution commitment problem Regime-types: Autocrats vs democrats Nationalism and world government Geographies of scale Chapter Illustrative empirics: introduction Unification vs balancing in contemporary Europe When anarchy is a rational choice: The case of the American civil war Chapter Alternative explanations for anarchy: The commercial peace hypothesis Global governance vs global government Born to kill? Chapter Conclusion: The paradox of world government References... 54

4 Chapter Introduction and research puzzle A central claim of neorealism is that the absence of world government has tragic implications for international politics, causing uncertainty, conflict, and war. 1 At the same time, realists tend to be opposed to the idea of world government, and rational explanations for why powerful states are reluctant to delegate military force when the absence of world government is assumed to cause tragedy and war is largely absent in realist literature. 2 Realist arguments for why international politics takes place in anarchy when great powers should have rational reasons to escape it (the paradox of anarchy) are underdeveloped and implicit, which indicates a gap in the explanatory power of neorealism. This gap motivates a theory-developing study that can provide a theoretically coherent solution to the paradox of anarchy. The paper contributes to international relations theory by developing a realist model departing from the assumption that anarchy is a rational choice rather than a deterministic structure. The paper considers a modified version of Fearon s realist two-state model which assumes that territorial greed (determined by regime-type, congruence between national borders and national identity and income inequality between states), military offense/defense balance (the probability of winning for the attacker given the levels of arms held by the defender) and geographic scale determine the cost of anarchy relative to world government. 3 The paper tests the validity of the model by illustrating how the independent variables affect the cost of anarchy relative to world government in hypothetical two-state scenarios where states can choose to attack unify or balance. The study proceeds by relaxing the two-state assumption to consider the impact of geographic scale on territorial greed and the offense/defense balance, as well as elucidating the logic of the model via empirical case illustrations. The paper suggests that when anarchy is costly due to offensive advantage and high levels of territorial greed, states rather take the chance of fighting to end future military competition than delegating military force to a world government that might be captured by poor states, which would imply redistribution of income from wealthy great powers. 1 See Waltz (1979) and Mearsheimer (2014). 2 Ibid 3 Fearon (2015). 1

5 When military defense dominates and territorial greed is low, world government becomes more achievable but anarchy less costly. Hence wealthy great powers are incentivized to balance in anarchy with a small military burden rather than delegating their military capabilities to a world government where they risk losing bargaining power gained from arming in anarchy. Furthermore, geographies of scale can plausibly explain why governments have evolved at the regional level but not the global level since military defense dominance increases with geographic scale, hence causing international anarchy to be more stable than regional anarchy. 1.2 Research questions and strategy The purpose of the study is to develop a theoretical model that can explain why rational states abstain from delegating their sovereignty to a world government, when anarchy is assumed by neorealist approaches to cause conflict and war. The study can hence be defined as theorydeveloping and attempts to eliminate a central theoretical puzzle in neorealism, defined in the study as the the paradox of anarchy. To solve the research puzzle the paper asks one explanatory and one descriptive research question: - If anarchy causes tragedy and war, then why are great powers reluctant to delegate their sovereignty to a world government? - What primary factors determine the cost of anarchy relative to world government? Blaikie emphasizes that it is important to distinguish between understanding (reason explanation) and explanation (causal explanation). Explanation refers to identifying the causal mechanisms that produce an empirical phenomenon while understanding refers to the motivation or account that social actors provide for their behavior. 4 The realist model presented in this paper is explanatory, and developed through deductive logic. The research strategy applied in the paper can be defined as a hybrid between deductive and retroductive strategies which are tailor-made for answering why questions, hence the most valid strategies for answering the research question of why powerful states choose to abstain from unifying when anarchy causes conflict and war. 5 4 Blaikie (2010) p, Ibid. 2

6 The deductive research strategy implies that the researcher presents a tentative idea, in this case, the idea that anarchy is a rational choice and not a structure which constrains state behavior deterministically. 6 Conclusions are then deduced from a set of axioms based on under what conditions the proposed hypothesis are expected to hold. In the present case, the cost of anarchy relative to world government as a function of military offense/defense balance, territorial greed and geographic scale. The realist model suggested in the paper is underpinned by the positivistic ontological assumption that social and political reality exists independently of our knowledge of it. Causes and consequences in the social world area assumed to resemble those of the natural world, thus social patterns are also natural phenomena, but since the actor preferences assumed by the model are not directly observable, the model developed in the paper draws upon a critical realist approach to social sciences. 7 The paper will not test the theoretical model on a single empirical case analyzed in depth, since the study is theory developing. Instead, it will utilize empirics mainly as illustrative examples, thus resembling elements of a retroductive research strategy, which seeks to capture underlying causal mechanisms in particular contexts. The first step in the retroductive research strategy is to provide an adequate description of the empirical regularity to be explained. In this case, the absence of world government. The second step is to identify contending mechanisms that can explain the regularity, which is done in the previous research and alternative explanation sections of the paper. 8 The paper considers a two-state model where states can choose to arm and attack, cooperate and divide gains from the issue at stake or delegate their sovereignty to a world government. The model highlights the similarities between constructing political order at the domestic and global level, the crucial difference being that preference heterogeneity and military defensive advantage is greater at the global level due to geographies of scale. The inquiry proceeds by illustrating how variation in military offense/defense balance, income inequality, regime types, congruency between national identity and state borders as well as geographies of scale, determines the cost of anarchy relative to world government. 6 Blaike (2010) p, Grix (2010) p, Blaike (2010) p, 87. 3

7 Chapter Previous research: introduction The chapter summarizes the central realist arguments for why great powers abstain from delegating military force, ranging from Waltz argument that there is no natural harmony of interests in the world to Hobbes argument that interstate competition is welfare improving for the subjects of the rulers. I argue that while there are arguments scattered across realist literature, they are largely implicit, informal and incoherent. The main flaw of previous realist arguments relating to the rationale for anarchy is that anarchy is assumed as a structure that deterministically forces states to compete for power, which leads to wasteful military spending and war. The chapter also provides a brief overview of historical state formation to highlight the impact of geographic scale on the cost of anarchy/world government. The chapter further discusses neo-liberal institutionalist explanations for interstate cooperation, and concludes that the concept of sovereignty costs which is a neoliberal explanation for why states do not want to delegate armed forces is indicating the importance of nationalism as an obstacle to cooperation. However, neo-liberal institutionalists do not explain explicitly why sovereignty is so important to states and do not confront the fact that states could avoid wasteful military spending but letting a global government guarantee their safety. 2.2 The virtues of anarchy While most realists have opposed the idea of world government, realists Morgenthau and Hertz argued in the 1950s that nuclear weapons made nation states defenseless, and called for the necessity of a global government, although none of them thought it would be enforceable. 9 While recognizing that world government is a logical solution to the dangers of anarchy, Morgenthau and Hertz fail to realize that precisely because nuclear weapons imply mutual assured destruction, great powers have a strong incentive to avoid fighting each other, hence anarchy becomes less rather than more dangerous. The model proposed in the paper assumes that nuclear weapons shifts the offense/defense balance in favor of defense, hence making anarchy more stable and less costly relative to world government. 9 Bell (2010) p,

8 Mearsheimer along with Hertz argue that anarchy implies a security dilemma, since states cannot gain in security without making other states less secure. Hence states are incentivized to acquire more and more power at others states expense, causing fierce security competition. 10 The puzzle is that if anarchy forces all states to compete intensely for power, rational states should anticipate fierce security competition and agree on a world government to eliminate the security dilemma. Furthermore, I argue that survival which is the principle goal of states according to offensive realism is not a precondition for all other possible ends since states could expand their consumption frontier by unifying and spend close to zero on arms just or keep down revolts from their subjects. 11 In Theory of international politics Waltz argued that what distinguishes the national from the international realm is that national governments possess a legitimate monopoly on violence, while the international realm is a self-help system. 12 Waltz compares the international selfhelp system to an oligopolistic market and argues that firms are constrained to strike a compromise between maximizing profit and minimizing the danger of being demised. The realist relative gains assumption which is one of the primary realist arguments for why international politics is competitive rather than cooperative appears to have its theoretical foundation in Waltz s argument that firms should always abstain from maximizing profit in order to maintain their market share. 13 However, Waltz model of market share maximizing firms was developed prior to the game theoretical idea that repeated interactions could allow firms to construct self-enforcing agreements that would sustain higher profits. 14 Powell argued in response that realism does not need a relative gains assumption since states in anarchy could arm to convert relative advantage into absolute gains. 15 Depending on military technology and costs for fighting, this possibility might prevent states from cooperating despite a shadow of the future (i.e., repeated interactions) Mearsheimer (2014) p, Mearsheimer (2014) p, Waltz (1979) p, Waltz (1979), p, Fearon (2015) p2. 15 Powell (1991) p, Fearon (2015) p2. 5

9 The paper departs from Powell s assumption that states seek absolute gains like firms on a market, but while Powell does not explain why states cannot unify to avoid the costs of anarchy, this paper shows how concerns over loosing relative bargaining power once military power has been delegated to a world government incentives states to choose anarchy. Waltz implicit argument for why states may not be able to unify in order to escape anarchy is that anarchy forces even good states that argue for world government to settle for balance of power as an unhappy alternative. 17 Waltz also argues that the same is true for domestic politics since men are products of the societies they live in, but why then has governments emerged at the local level and not at the global level? Waltz implicit answer is that because there is no natural harmony of interests among states and no global leviathan to prevent war, there is a latent risk that states will use force to achieve their apparent issue indivisible ends. 18 If there are no natural harmony of interest in anarchy, the logical negation of the argument would be that there would be a natural harmony of interest under hierarchy and this is indeed what is assumed to be true at the domestic level in Waltz realist analysis. 19 The national interest would be impossible to define without assuming that citizens have mutual interests that aggregates into a national interest which is defined in terms of power. If all states have the same rational objectives under anarchy, then it is puzzling why great powers wouldn t have incentives to federate to avoid the violent clash of interests in anarchy. Waltz elaborates on the undesirability of world government by arguing that: It would be an invitation to civil war since a global central authority could not mobilize sufficient resources to regulate and manage its parts. 20 A more explicit formalization of Waltz argument would be that states may fear that global governmental institutions might not be robust against capture, which together with fear of redistribution of resources determines the cost of world government in the RCR (rational choice realist) model suggested in this paper. Global civil war would not occur if there were no conflicting preferences however, hence income inequality which causes redistribution disputes and nationalism which causes cultural and territorial disputes are central independent variables in the RCR model. 17 Waltz (1954) p, Ibid 19 Ibid 20 Waltz (1979), p,

10 That nationalism reinforces anarchy tend to be an implicit assumption in realist analysis but rational explanations for why states are inherently nationalistic are largely absent. Waltz argues that states like people are insecure in proportion to the extent of their freedom. 21 What states want freedom from is not explained however, and while Mearsheimer is more explicit about the role of nationalism in international relations by arguing that nationalism cases glorification of the state, he does not provide rational arguments for why states should prefer sovereignty in an anarchic system where costly arming is required to remain sovereign Territorial competition and the cost of anarchy One of the most convincing neorealist arguments for why anarchy would prevent global unification is provided by Jervis who interpret international cooperation as a commitment problem where each state has an incentive to defect because: a) States currently supporting the status quo might get dissatisfied in the future. b) States tend to seek control over resources beyond their borders to assure that the necessary supplies will continue to flow during wartime. 23 Jervis argues that if there would to be an international authority (world government) that could guarantee access to external resources, the motive for controlling territory would disappear. 24 Thus, Jervis highlights that if a world government existed, the resource-problem would be eliminated but he doesn t really explain why rational great powers wouldn t prefer to delegate force to a global leviathan to end the commitment problems arising from anarchy. Jervis appear to assume by default that states find themselves in a security dilemma in which they can t trust each other s intentions. The realist argument that states need to secure external resources provides a valuable concept that will be incorporated in the RCR model, where territorial greed is assumed to have a decisive impact on arms levels and cost of anarchy relative to world government. 21 Waltz (1979), p, Mearsheimer (2001) p, Jervis (1978) p, Ibid. 7

11 2.4 Hobbesian contract between men but not between sovereigns In contrast to neorealism, classical realism assumes that it is not the structure of the international system but human nature that is the primary cause of international conflict. 25 However, classical realist Hobbes implicitly provides a tentative answer to neorealist paradox of anarchy by arguing that while domestic anarchy implies misery, anarchy between sovereign authorities forces rulers to uphold the industries of their subjects in order to compete successfully with their rivals. 26 Hobbes argue that men are inclined to domestic peace by the passions of fear of death and desire for commodious necessities. 27 The puzzle is that while the sovereign s subjects may have their wealth and security protected by the sovereign, the sovereigns themselves should have rational reasons to eliminate the fear of death in the international realm by surrendering their military force to a global sovereign, hence converting their military burden to wealth and enjoy the protection of a global leviathan. 2.5 International cooperation and institutional design Neoliberal institutionalism, representing the main theoretical competitor to realism within the rationalist spectra, has provided extensive theoretical arguments for why world unification might be prevented despite costs to anarchy. Koremenos, Lipson and Snidal asks what prevents international cooperation if it is always possible through repeated games (Folk theorem) implying that international anarchy should be overcome after repeated interactions between states. 28 Koremenos et al argues that the Folk theorem has been relying simple 2x2 prisoner s dilemma game where there is only one equilibria outcome while in reality there are often many. 29 Morrow, Krasner and Fearon have also shown how distributional differences can undermine cooperation, and based on mainly the work of Fearon, income-inequality will figure as one of the primary variables determining the cost of anarchy relative to world government. Koremenos et al concludes that distribution problems are as significant of an obstacle to international cooperation as enforcement problems Mearsheimer (2001) p, Hobbes (2004) p, Ibid 28 Koremenos, Lipson and Snidal (2001) p, Koremenos, Lipson and Snidal (2001), p, Ibid 8

12 Uncertainty and noise (observe other actors actions clearly) is another problem, because states are often reluctant to disclose vital information that could make them more vulnerable. 31 Koremenos et al also argue that large number of players complicates cooperation since some states are richer and more powerful than others. 32 This paper revises the argument, since the premise appears to be that more players implies more diversity and more diversity implies less cooperation, which motivates nationalism, income inequality and regime-types as central variables determining the cost of anarchy relative to world government in the RCR model. Furthermore, Koremenos et al. argue that concerns about sovereignty explains why states abstain from delegating strong coercive capacities in spite of possible gains that could be made from centralization, but it is unclear why states find sovereignty critically important when enforcement problems and transaction costs can be reduced by agreeing on maximum global centralization of bargaining (world government). 33 The RCR model provides firmer justification for the sovereignty cost argument by emphasizing the importance of nationalism as an obstacle to cooperation in the military realm, where nationalism is treated as a rational tool for state leaders to enhance national unity and fighting efficiency, hence reducing the cost of domestic hierarchy and international anarchy. Oye has argued that states can arrive at mutually beneficiary cooperative equilibria by prolonging the shadow of the future. 34 Schelling and Axelrod suggest tactics of decomposition over time to lengthen the shadow of the future, thus reducing the risk of defection by splitting up for example a deal to delegate all military force to a supranational world government into several periods of bargaining where some military capability is delegated step by step in exchange for reduced military burdens and influence over the world government. 35 However, Fearon has argued that the shadow of the future does not eliminate the war constraint, which he defines as the incentive for state A to attack and conquer state B if state B would to disarm, since state A then could arm a little and conquer B at very low costs relative to the benefit of resource extraction from state B Koremenos, Lipson and Snidal (2001), p, Ibid 33 Koremenos, Lipson and Snidal (2001) p, Oye (1985) p, Ibid 36 Fearon (2015) p, 10. 9

13 Thus, the shadow of the future might not be prolonged such that states would disarms completely, which means that if states would delegate military force to a world government in repeated games, incentives to defect and attack a delegator temporally weakened by disarmament could prevent a completion of the delegation process and risk war and mutual defection. Oye makes the fundamental assumption that because there is no central authority to impose limits on the pursuit of sovereign interests, anarchy gives rise to diverse outcomes, without explicitly recognizing that the absence of world government is an equilibria outcome of heterogeneous national interests. Oye also appears to make a tautological argument for why states do not want to delegate their sovereignty by assuming that the absence of world government precludes that states would cede ultimate control over their conduct to a supranational sovereign. 37 Hence, Oye fails to answer the question of how world government would solve the commitment problems arising from diverse national interests such that it would eliminate tragedies of anarchy The historical evolution of international anarchy The central premise of the rational choice realist approach is that anarchy is an outcome of rational calculus. However, states evolved bottom-up, not top down, which motivates a historical summary of the evolution of the state system which constitutes contemporary international relations. Thus far, the paper has been posing the question of why world government has not been enforced by contract, but world history shows that state formation has been largely driven by coercive bargaining and war. 39 The predatory approach to state formation advocated by Tilly suggests that states in medieval Europe built fearsome coercive means to deprive civilian populations of access to those means, and relied on capitalists to reorganize coercion. 40 Tilly does not make explicit assumptions about exactly what the fundamental goal of rulers are, whether they are assumed to be security seekers or greedy territorial expanders, but instead assumes that men who control concentrated means of coercion wants to extend them over population and resources over which they wield power Oye (1985) p, Ibid 39 Tilly (1992) p, Ibid 41 Tilly (1992) p,

14 Furthermore, Tilly implicitly makes the offensive neorealist assumption that whenever they encounter other rulers with no comparable control of coercion, they have an incentive to conquer, and when faced with a comparable strong rival they make war. 42 Rulers are constrained with significant limits to its range of control, which causes either military defeat or fragmentation of control, thus hinting that international anarchy is a product of a lack of global concentration of the means of coercion. Because of a lack of external control Tilly argues that rulers historically settled for a combination of conquest, protection against powerful rivals and cooperation with neighbors. 43 Since the dawn of civilization, rulers have attempted to establish a secure territory within which they could enjoy returns from coercion and a fortified buffer zone to protect that territory. When the effort was successful the buffer zone became a secure area, which incentivized the ruler to acquire additional buffer zones surrounding the old. Since all rulers where assumed to follow the same logic, adjacent rulers inevitable engaged in military conflict. 44 Small kingdoms evolved into nation states when armies were nationalized and specialized after the 1750s through a process of bargaining out access to households and enterprises, hence sweeping out autonomous intermediaries. 45 Hence, coercion and asymmetric bargaining interacted to concentrate capital and coercion. The paper argues that since it requires more offensive capability to control a large geographic area than a small one, geographic scale is an important factor explaining why hierarchy emerged at the local level but not the global level. On the other hand, Tilly highlights that even if states had significant bargaining leverage over its subjects, rulers often bargained peacefully with revolting subject to avoid the costs of coercion. 46 Moreover, Tilly s emphasis on the effectiveness of coercion largely neglects the wars inefficiency puzzle. War is costly and rational rulers should have an incentive to federate with their competitors, so to spend just enough on arms to keep down revolts from their subjects and protect wealth from capture. Predatory theories of state formation which are consistent with a fundamentally realist understanding of state behavior may explain how anarchy evolved into local hierarchy trough coercion and asymmetrical bargaining, but fails to explain why rational rulers shouldn t have an incentive to federate at the global level to spend just enough on arms to prevent revolts and protect wealth from capture. 42 Tilly (1992) p, Ibid 44 Tilly (1992) p, Tilly (1992) p, Tilly (1992) p,

15 Chapter Method and operationalization: introduction The paper utilizes Frieden s method of deducing actor preferences to generate predictions about how observable variations in state properties should change state preferences over international political order (anarchy or world government as comparative static) given an underlying profit motive derived from microeconomic assumptions about firm. The congruence method will be applied to test if the predictions of the RCR model are congruent with empirical outcomes in illustrative cases, but since the paper is theory developing, empirics will mainly be utilized to illustrate the logic of the RCR model. 3.2 The congruence method The congruence method implies that the researcher ascertains the value of the independent variable and predicts, based on the theory used, what outcome of the dependent variable that should follow. 47 George and Bennet emphasize that the congruence method is useful in studies that work with deductive theories that black-box decision-making and strategic interaction, in this case, state leaders choosing between anarchy and world government in two-state game scenarios. 48 Hypothetical two-state scenarios can be thought of as artificial small-n case studies that illustrates macro-political dynamics by analyzing strategic interactions between two states. The first step in a theory developing study that black-box internal decision-making processes is to formulate a formal version of the deductive theory employed. 49 The formal model developed in this paper which will be assessed in detail in chapter 4 can be defined as a rational choice realist model where two states can choose to unify to eliminate arms spending, balance to maintain deterrence or fight in hopes of eliminating future military competition. Costs and benefits of unification, balancing and war are determined by the military offense-defense balance and territorial greed. 47 George and Bennet (2005) p, George and Bennet (2005) p, Ibid 12

16 The second step is to identify historical cases that enables refinement of the theory s explanatory power. 50 In this case, the absence of an outcome is the dependent variable to be explained, but since the outcome (world government) has never occurred, there is no empirical variance in the dependent variable unless anarchy is treated as a variable varying between less or more anarchy rather than pure anarchy and world government. However, since the paper assumes anarchy (absence of world government) to be a rational choice, the dependent variable (international order) can vary between anarchy and world government since great powers could potentially choose to unify. A third and last step in the application of the congruence method is to test how congruent the predictions of the RCR model are with empirical outcomes, which will be done by assessing the German-French plans for a European defense union which if realized would be a case of peaceful unification in the RCR model, while the American civil war will be analyzed as a case of where regional world government dissolved into temporal anarchy Deducing actor preferences Friedens method of deducing actor preferences provides a plausible tradeoff between methods of observation that tend to confuse preferences with outcomes and methods of assumption that neglects how variation in state properties affect preferences. 52 Deducing actor preferences implies assuming a fundamental exogenous preference derived from previous theory and then determine preferences over specific issues given identifiable characteristics of actors and their environment. 53 Based on the realist firm-state analogy, The RCR model assumes that state leaders have a fundamental preference for maximizing consumption, but whether state leaders will prefer anarchy or world government is determined by variation in observable comparative statics (regime-types, relative income, congruency between the national and political unit) and structural factors (military technology and geographic setting). Frieden argue that the deductive method is analytically valuable because comparative statics are more readily observed than preferences, hence the RCR model can make predictions about how state preferences are likely to change given variation in observable state properties and the military offense-defense balance George and Bennet (2005) p, George and Bennet (2005) p, Frieden (1999) p, Frieden (1999), p, Frieden (1999), p,

17 Frieden highlights the importance of keeping preferences separate from the characteristics of strategic settings, or else the causal role of actor interests and their environment cannot be distinguished. 55 In any given setting, an actor prefers certain outcomes over others and pursue strategies to achieve the most preferred outcome. In this paper, states are assumed to be the primary actors in international politics and their principal goal is consumption maximization of material and symbolic goods. State leaders can choose three different strategies to obtain this goal. Conquest (outcome: world government), power balancing (outcome: anarchy), or delegation of force to supranational institutions (outcome: world government). Frieden highlights that blending national preferences, strategies and the environment into a single explanatory factor makes it difficult to know how outcomes are affected by international power relations, national interests and features of the environment. 56 Frieden argues that the two main analytical problems related to the blending of state interests, strategies and strategic setting are sins of commission which means that it is an analytical mistake to assume that variation in outcomes is solely due to variation in preferences, and sins of omission which implies that variation in outcome has nothing to do with variation in preferences. 57 The sin of omission can be avoided by holding environmental factors constant while letting comparative statics vary, which is done in the RCR model by assuming two-state scenarios where geographic scale is held constant, while relative incomes, offense/defense balance, regime types and congruency between national identity and borders vary. Likewise, the sin of commission can be avoided by holding comparative statics (state properties) constant while varying environmental factors (geographic scale in the RCR model). 55 Frieden (1999) p, Frieden (1999) p, Ibid. 14

18 3.4 Delimitations, reliability and validity Realist and game-theoretical approaches can usually only make very general probabilistic predictions and George and Bennet argue that probabilistic prediction is not adequate for assuming a causal relationship unless other outcome explanations are assessed and eliminated. 58 Thus, it is important to assess whether the independent variable is a necessary or sufficient condition for the outcome to occur. In the RCR model, each independent variable is a necessary but not sufficient condition for an outcome to occur. George and Bennet claim that structural realism and game-theories lack operationalization and even when operationalized may fail to identify satisfactory accounts of causal mechanisms but I argue that causal mechanisms are implicit in the internal logic of the deductive realist theoretical model and if the model generates successful predictions it needs no further explanation or demonstration. 59 The RCR model has high external validity (generalizability) due to its causal theoretical mechanisms, but its conceptual validity is lower due to that it relies on a series of assumptions that can be disputed. George and Bennet mention that statistical studies run the risk of conceptual stretching, which is also true for grand theoretical models, hence the simple dichotomy between anarchy and world government may be disputed. 60 The model also faces an equifinality problem since the absence of world government could be attributed to human nature (emotional fear of the intentions of other states due to anarchy) or lack of a common enemy which would prevent unification, given that power balancing is the only rational motive for cooperation. 61 The paper will deal with the equifinality problem by discussing alternative realist and neoliberal explanations for anarchy in chapter 7. The issue of conceptual validity also implies reliability problems. Small changes in the assumed institutional framework of a potential world government could imply large fluctuations in the cost of anarchy. However, the study does not attempt to provide and exact measurement of the cost of anarchy relative to world government, but rather provides a logically feasible realist model that can explain why state leaders prefer anarchy over unification. 58 George and Bennet (2005) p, George and Bennet (2005) p, George and Bennet (2005) p, George and Bennet (2005) p,

19 Even if the potential costs of world government cannot be empirically measured since world politics has always been anarchic, Fearon has shown that all levels of governance share the same basic commitment problem. 62 In the same way as an ethnic majority in a civil war would not be able to credibly commit not to exploit the greater bargaining leverage it would gain over ethnic minorities if a new state would be consolidated, a large poor state would have a difficult time credibly committing not exploit the greater bargaining leverage it would gain if it unified with a richer smaller state. Hence, the small wealthy state has rational reasons to abstain from unification to avoid redistribution of income. Since domestic and global politics share the same basic commitment problems, two-state scenarios are useful as illustrative cases of how competing nationalisms, differences in relative income, variation in regime-types and military offense-defense balance can affect the cost of world government relative to anarchy. Measuring the cost of anarchy is less problematic since military spending provides a plausible empirical indicator of how much states have to spend on the military to maintain deterrence. However, there are balance of power dynamics that affect the validity and reliability of military spending as a rough estimate of the cost of anarchy relative to world government. Fearon argues that there is a tendency for the smaller state in a dyadic rivalry to make a larger military effort relative to its resource base, but there are two have to offsetting effects for the wealthier state in the dyad. 63 The wealthier state can spend a smaller fraction of GDP on the military to sustain deterrence, but also has to spend more to keep the poorer state from invading it since war is partly caused by undefended wealth. 64 These offsetting effects imply that m (offensive advantage) and µ-c (territorial greed-other war costs) has to be large in order for differences in relative resources to decisively impact military burdens. 65 Given that relative material resources has implications for military spending never the less indicates that arms levels are not solely dependent on m and µ-c, but also relative resources. 62 Fearon (1995) p, Fearon (2015) p, Ibid. 65 Fearon (2015) p,

20 Chapter Theoretical framework: Introduction This chapter introduces a modified version of Fearon s two-state baseline model where a third possible equilibrium outcome and a constraint tied to that outcome is added. In the original twostate model, anarchy is taken as a structural given where states can choose between balancing and attacking in successive periods, and because states inevitably must arm to deter each other (the war constraint), full cooperation in anarchy is not possible. The modified model, denoted as the RCR (rational choice realist) model assumes that anarchy is not a structural constraint which produces deterministic outcomes but a rational choice that state leaders make in preference to world government. The RCR-model extends the implications of Fearon s twostate model by introducing a delegation constraint, which explains why states abstain from peaceful unification. The delegation constrain holds that states can eliminate the war constraint by unifying and avoid costly deterrence, but rich powerful states may lose bargaining power by disarming, since they might be forced to redistribute material goods to poor states if the world government is democratic. However, if the world government is autocratic with insufficient checks and balances, the war constraint holds, since the world government may become tyrannical and coerce its subjects. The RCR-model extends the implications of the Fearonian model by showing how military defense dominance (determined by military technology) and low levels of territorial greed (determined by the congruency between national identities and borders, income inequality and regime-types) makes anarchy stable and world government less urgent, while offense dominance and high levels of territorial greed makes anarchy costly, but also world government fundamentally unstable. Hence states rather take the chance of going to war to eliminate future military competition than unifying with greedy competitors that might grab supranational military forces to impose their cultural preferences on other states or push for redistribution of resources. Moreover, the RCR model adds the independent variable geographic scale which accounts for why anarchy is more stable at the global than the regional level, hence explaining why there are national governments, but no world government. Furthermore, the model assumes democracy to be an institutional constraint, not an ideational factor that fundamentally changes the preferences of state leaders. 17

21 Likewise, nationalism is treated as an instrumental variable that leaders promote to enhance the fighting morale of soldiers, hence providing rational arguments for why nationalism and regime-types should figure in a formal realist model. 4.2 Fundamental assumptions Neorealist approaches typically depart from the assumption that the principal goal of states is survival and to survive states relentlessly seek security by maximizing their relative share of world power. 66 I argue that the survival motive is logically problematic since it assumes that states exist for self-preservation purposes, based on a state-individual analogy. 67 While individuals have to survive to pursue other ends, state survival is not a precondition for other ends that individuals may pursue since state unification could enhance per capita consumption. The paper argues that a more feasible fundamental principle goal of state leaders is consumption maximization of symbolic and material goods. The consumption maximization assumption solves the puzzle arising from the common realist assumption that the primary motive of states is to survive, and yet state autonomy may block economic gains from unification. Under anarchy, the consumption frontier for any state is suboptimal since a fraction of the potential consumption must be spent on arms to deter potential aggression and gain bargaining power. The paper therefore assumes arms spending to be an inefficiency cost that states want to minimize. Mearsheimer argues that non-security goals sometimes complement the hunt for relative power and that greater economic prosperity has significant security-implications due to the long-run harmony between wealth and power. 68 The preference order should hence be reversed, since latent power (wealth) is a precondition for real power (military power). Mearsheimer agrees that states cannot build powerful militaries without the money and technology to equip them but views wealth as means to the end of survival, not survival as means to the end of consumption. 69 The empirical feasibility of the consumption maximization motive can be exemplified by the US post war grand strategy to secure open markets in Europe and Asia, since the CIA concluded in 1947 that a possible collapse of the western European market was the greatest security threat to the US Mearsheimer (2014) p, Fearon (2015) p, Mearsheimer (2014) p, Mearsheimer (2014) p, Ikenbeery (2011) p,

22 Furthermore, Roosevelt concluded that the US could not continue to thrive a superpower after the Second World War without access to markets outside the western hemisphere that could fuel post war growth. 71 Another core assumption in structural realism is that states are concerned with relative gains which implies that cooperation becomes difficult if states are unable to divide gains such that relative bargaining power will not change in the future. 72 If states seek relative gains the statefirm analogy does not hold however, since modern microeconomics assumes that firms are profit maximizers, not market share maximizers. The RCR model therefor assumes Powell s proposition that states in anarchy seek relative military advantage to maximize absolute gains. 73 Neorealist approaches typically assume states to be unitary actors. 74 The RCR model assumes state leaders, not states, to be the primary actors in international politics, which allows the RCR model to captures how regime-type constrains the ability of state leaders to maximize their consumption. The regime-type constraint implies that autocrats are less obligated to share benefits of territorial conquest among subjects than democratically elected leaders since citizens in a democracy have more bargaining power. Democrats and autocrats have the same basic objective, to maximize consumption, but because they face different institutional constraints, the paper makes the simplified assumption that democracies spend resources not devoted to arms on public goods while autocracies are assumed to be rent maximizers. The RCR model also assumes that state leaders are boundedly relational due to imperfect information which incentivizes them to fear that other state leaders will not credibly commit to not take advantage of shifts in bargaining power between states, if they would to unify into a world government. 71 Ikenbeery (2011) p, Fearon (1998) p, Powell (1991) p, See Waltz (1979). 19

23 4.3 Theoretical model The basic Fearonian model with no variation in relative power assumes that two states choose arms levels simultaneously in successive periods, and calculate whether to attack or continue to bargaining over issue y (international order in this case). 75 In the RCR model, states do not only choose between anarchy or world government through acquisition, but can also choose to peacefully unify. Either war breaks out and state A conquers the territory of state B (world government through acquisition) or both states agree on zero arms spending and delegates their military forces to a world government, hence peaceful unification results. In both cases, anarchy is overcome and the game ends. If the balance of power (arms levels) is stable, anarchy results. The RCR model considers the impact of geographic scale when relaxing the two-state scenario and allows for a third possible outcome, world government by institution as well as introducing a delegation constraint (the constraint state leaders face when delegating military force to supranational institutions, losing potential bargaining power but gain potential consumption by reducing wasteful arms spending). 76 While the RCR model assumes that state leaders are fundamentally rational, imperfect information incentivizes them to fear that their adversaries will be unable to credibly commit to not attempt to capture world governmental military forces or push for redistribution of economic resources from rich to poor regions. The paper argues that when the two-state assumption is relaxed, geographic scale makes the equilibria outcome war by acquisition unlikely due to that defensive advantage increases with geographic scale. The model also predicts that if there is a third balancer in the game that would intervene on behalf of state A if state B would to attack, cooperation between state A and B becomes easier, defense spending lower and anarchy less costly. If a third state threatens the security and prosperity of both state A and B, they are likely to form a coalition to balance against state C. The paper will first present the basic intuition of the Fearonian two-state model and proceed with implications of relaxing the two-state scenario. 75 Fearon (2015) p, 6. 20