David Baldwin. Handbook of International Relations. Editor: Walter Carlsnaes, Thomas Risse, Beth A Simmons. Sage Publication. 2002.
Most definitions of politics involve power. Most international interactions are political or have ramifications for politics. Thus, it is not surprising that power has been prominent in discussions of international interaction from Thucydides to the present day. The long history of discussions of the role of power in international relations, however, has failed to generate much agreement. Scholars disagree not only with respect to the role of power but also with respect to the nature of power. One scholar notes that power is a key concept in realist theories of international politics, while conceding that ‘its proper definition remains a matter of controversy’ (Waltz, 1986: 333). Robert Gilpin describes the concept of power as ‘one of the most troublesome in the field of international relations’ (1981: 13) and suggests that the ‘number and variety of definitions should be an embarrassment to political scientists’ (1975: 24). There is, however, widespread consensus among international relations scholars on both the necessity of addressing the role of power in international interactions and the unsatisfactory state of knowledge about this topic (Guzzini, 2000). Although it is often useful to distinguish among such power terms as power, influence, control, coercion, force, persuasion, deterrence, compellence, inducement and so on, it is possible to identify common elements underlying all such terms. Robert A. Dahl (1957) has suggested that underlying most such terms is the basic intuitive notion of A causing B to do something that B otherwise would not have done. (In the discussion that follows, ‘A’ refers to the actor having or exercising influence; while ‘B’ refers to the actor being influenced.) Although alternative definitions of power abound, none rivals this one in widespread acceptability. In the following discussion, the term ‘power’ will be used in a broad generic sense that is interchangeable with such terms as ‘influence’ or ‘control’ unless otherwise indicated. This usage is not intended to deny the validity or the utility of distinguishing among such terms for other purposes.
Power and the Study of International Politics
International politics has been defined in terms of influencing ‘major groups in the world so as to advance the purposes of some against the opposition of others’ (Wright, 1955: 130). Although the term ‘power politics’ has unsavory connotations for some, such a definition implies that the term is redundant (Carr, 1946; Morgenthau, 1960; Sprout and Sprout, 1945; Spykman, 1942; Wright, 1955). From this perspective, all politics is power politics in the sense that all politics involves power. This is not to say that politics is only about power.
Traditionally, the study of international politics assumed the existence of national states with conflicting policies, placing a high value on maintaining their independence, and relying primarily on military force. The states with the most military power were designated ‘Great Powers,’ and the ‘game’ of international politics was ‘played’ primarily by them (Sprout and Sprout, 1945, 1962; Wight, 1946). Noting that only a few states possessed the military capabilities to support their foreign policies effectively, an influential text in the 1930s averred that ‘these alone constitute the Great Powers’ (Simonds and Emeny, 1937: 28).
In the eighteenth century, ‘the power of individual states was conceived to be susceptible of measurement by certain well-defined factors’ (Gulick, 1955: 24), including population, territory, wealth, armies and navies. In the ensuing years, this approach evolved into the ‘elements of national power’ approach to power analysis reflected in Hans J. Morgenthau’s influential textbook Politics Among Nations ( 1960 see also Sprout and Sprout, 1945).
States were depicted as seeking to maximize power relative to each other, thus producing a ‘balance of power’ or as seeking to produce a balance of power (Claude, 1962; Gulick, 1955; Haas, 1953; Morgenthau  1960). Each version of balance of power theory shared the assumption that it was possible to add up the various elements of national power, sometimes called ‘power resources’ or ‘capabilities,’ in order to calculate the power distribution among the Great Powers. A modern version of this approach is found in Kenneth N. Waltz’s Theory of International Politics (1979).
The Power Analysis Revolution
The ‘elements of national power’ approach depicted power as a possession or property of states. This approach was challenged during the last half of the twentieth century by the ‘relational power’ approach, developed by scholars working in several disciplines, including psychology, philosophy, sociology, economics and political science (Baldwin, 1989; Barry, 1976; Cartwright, 1965; Dahl, 1957, [1963, 1984] 1991; 1968; Frey, 1971, 1985, 1989; Harsanyi, 1962; Nagel, 1975; Oppenheim, 1981; Simon, 1957; Tedeschi and Bonoma, 1972). Some would regard the publication of Power and Society by Harold Lasswell and Abraham Kaplan (1950) as the watershed between the old ‘power-as-resources’ approach and the new ‘relational power’ approach, which developed the idea of power as a type of causation. This causal notion conceives of power as a relationship in which the behavior of actor A at least partially causes a change in the behavior of actor B. ‘Behavior’ in this context need not be defined narrowly, but may be understood broadly to include beliefs, attitudes, opinions, expectations, emotions and/or predispositions to act. In this view, power is an actual or potential relationship between two or more actors (persons, states, groups, etc.), rather than a property of any one of them.
The shift from a property concept of power to a relational one constituted a revolution in power analysis. Despite the ancient origins of the study of power, Dahl maintains that ‘the systematic empirical study of power relations is remarkably new’ (1968: 414). He attributes the ‘considerable improvement in the clarity’ of power concepts to the fact that ‘the last several decades have probably witnessed more systematic efforts to tie down these concepts than have the previous millennia of political thought’ (Dahl, [1963, 1984] 1991: 27–8).
Dimensions of Power
Power was no longer viewed as monolithic and unidimensional, but rather as multidimensional. This allowed for the possibility that power could increase on one dimension while simultaneously decreasing on another. Among the more important dimensions of power were the following.
Scope Scope refers to the aspect of B’s behavior affected by A. This calls attention to the possibility that an actor’s power may vary from one issue to another. Thus, a country like Japan may have more influence with respect to economic issues than with respect to military issues.
Domain The domain of an actor’s power refers to the number of other actors subject to its influence. In other words, how big is B; or how many Bs are there? Thus, a state may have a great deal of influence in one region of the world, while having little or no influence in other parts of the world.
Weight The weight of an actor’s power refers to the probability that B’s behavior is or could be affected by A (Dahl, 1957; see also Deutsch, 1988; Lasswell and Kaplan, 1950). Thus, a country that has only a 30 per cent chance of achieving its aims in trade negotiations is less powerful than one with a 90 per cent chance, ceteris paribus. This dimension could also be labeled the ‘reliability’ of A’s influence.
Costs Both the costs to A and the costs to B are relevant to assessing influence (Baldwin, 1989; Barry, 1976; Dahl, 1968; Harsanyi, 1962; Schelling, 1984: 268–90). Is it costly or cheap for A to influence B? Is it costly or cheap for B to comply with A’s demands? Some have suggested that more power should be attributed to an actor that can exercise influence cheaply than to one for whom it is costly (Harsanyi, 1962). If A can get B to do something that is costly for B, some would contend that this is indicative of more power than if A can only get B to do things that are cheap for B. Even if A is unable to get B to comply with its demands, it may be able to impose costs on B for non-compliance. Some have argued that this should be viewed as a kind of power (Baldwin, 1985; Harsanyi, 1962; Schelling, 1984: 268–90).
Means There are many means of exercising influence and many ways to categorize such means. One scheme (Baldwin, 1985) for classifying the means of influence in international relations includes the following categories:
- Symbolic means. This would include appeals to normative symbols as well as the provision of information. Thus one country might influence another either by reminding them that slavery is bad or by informing them that AIDS is caused by HIV.
- Economic means. Augmenting or reducing the goods or services available to other countries has a long history in world politics.
- Military means. Actual or threatened military force has received more attention than any other means in international relations.
- Diplomatic means. Diplomacy includes a wide array of practices, including representation and negotiation.
Which dimensions of power should be specified for meaningful scholarly communication? There is no single right answer to this question. The causal concept of power, however, does imply a minimum set of specifications. The point is well put by Jack Nagel (1975: 14):
Anyone who employs a causal concept of power must specify domain and scope. To say ‘X has power’ may seem sensible, but to say ‘X causes’ or ‘X can cause’ is nonsense. Causation implies an X and a Y–a cause and an effect. If power is causation, one must state the outcome caused. Stipulating domain and scope answers the question ‘Power over what?’
The idea that a meaningful specification of a power relationship must include scope and domain is widely shared by power analysts committed to social scientific inquiry (Barry, 1976; Dahl, 1991, 1968; Deutsch,  1988; Frey, 1971, 1989; Lasswell and Kaplan, 1950).
The multidimensional nature of power makes it difficult to add up the various dimensions in order to arrive at some overall estimate of an actor’s power. Although there are some similarities between political power and purchasing power (Baldwin, 1989), one important difference is the lack of a standardized measuring rod for the former. Whereas money can be used to measure purchasing power, there is no comparable standard of value in terms of which to add up the various dimensions of power so as to arrive at an overall total. For this reason, estimates of an actors ‘overall power’ are likely to be controversial.
Faces of Power?
One of the most famous debates in the literature on power during the last half of the twentieth century is known as the ‘Faces of Power’ debate (Bachrach and Baratz, 1962; Isaac, 1987; Lukes, 1974). Although this debate clarified some important points, its significance for international relations is sometimes misunderstood. The first issue was whether control over agendas could be accommodated. And the second issue was whether control over the desires and thoughts of others could be taken into account. Whatever the shortcomings of various empirical analyses of power, it should be noted that the basic causal concept of power discussed earlier can easily accommodate either of these phenomena. A can cause B to do something that B would otherwise not do by controlling B’s agenda (that is, B’s options) or by affecting B’s preferences, desires and thoughts. Brainwashing can be analyzed with the conventional relational concept of power. One does not need to reconceptualize power in order to treat such matters.
International Power Analysis
Although many political scientists have contributed to the power analysis revolution during the past fifty years, very few have been students of international relations (Baldwin, 1971b; Singer, 1963). Harold and Margaret Sprout, who had been proponents of the elements of national power approach in their early work (Sprout and Sprout, 1945), later repudiated that approach and were among the first international relations scholars to call for incorporation of the relational power approach into the study of international politics (Sprout and Sprout, 1956, 1962, 1965). Despite the efforts of the Sprouts and others, however, the elements of national power approach is still deeply embedded in the international relations literature (Waltz, 1979). This situation has given rise to several problems in the analysis of power in the international arena, some of which are discussed below.
The Potential Power Problem
The elements of national power approach to power analysis is a variant of the power-as-resources approach. In this approach, power resources are treated as if they were power itself. One problem with this approach is that what functions as a power asset in one situation may be a power liability in a different situation. Planes loaded with nuclear bombs may be worse than useless in a situation calling for planes with conventional weapons with insufficient time to unload the nuclear weapons and reload the planes with conventional ones. And the same stockpile of arms that is useful for deterring one country may trigger an arms race with another. Similarly, what constitutes a ‘good hand’ in card games depends on whether one is playing poker or bridge. Discussions of the capabilities of states that fail to designate or imply a framework of assumptions about who is trying (or might try) to get whom to do what are comparable to discussions of what constitutes a good hand in cards without specifying which game is to be played. The Sprouts called this set of assumptions a ‘policy-contingency framework’ (1965, 1971). Focusing on the capabilities of states is simply a way of drawing attention to their potential power. It makes no more sense to talk about state capabilities in general than to talk about state power without (explicitly or implicitly) specifying scope and domain. If one wants to estimate the potential power of Guatemala, it helps to know, nay, it is imperative to know whether it concerns a border dispute with El Salvador or a trade agreement with the United States.
Although it is sometimes suggested that insistence on specification of the scope and domain of potential power relationships makes prediction and or generalization nearly impossible (Guzzini, 2000; Keohane, 1986), this is not true. Specification of scope and domain (or policy-contingency frameworks) need not imply atheoretical empiricism. Policy-contingency frameworks may be defined more or less broadly to suit the purpose of the analyst. As Nagel (1975: 14) observes, ‘domain and scope need not be particularistic or unique. Depending on one’s purpose and the limits imposed by reality, the outcome class may contain a few similar members or many diverse elements.’ It is, of course, possible to make predictions or generalize about the potential power of Guatemala (or similar states) without reference to Guatemala’s goals and without reference to the goals or capabilities of other states; but it is not clear why one would want to do so.
Power resources are the raw materials out of which power relationships are forged. Although it might seem that the predictive value of power resource inventories is impaired by insistence on prior specification of scope and domain, the opposite is true. The accuracy of one’s estimate of whether an architect has adequate raw materials to complete his or her project is likely to improve if one first ascertains whether the architect plans to build a birdhouse or a cathedral.
The Fungibility Problem
‘Fungibility’ refers to the ease with which power resources useful in one issue-area can be used in other issue-areas. Money in a market economy is the prototypical fungible resource. Indeed, fungibility (that is, liquidity) is one of the defining characteristics of money. In a market economy one does not usually need to specify the scope or domain of the purchasing power of money because the same euro (yen, dollar, etc.) can be used to buy a car, a meal, or a book.
It is sometimes suggested that power plays the same role in international politics that money does in a market economy (Deutsch,  1988; Wolfers, 1962). Political power resources, of course, do vary in degree of fungibility. Money, time and information tend to be more fungible than most other power resources in that they are useful in many different issue-areas. To the extent that the power–money analogy leads to ignoring the need to specify scope and domain, however, it can be quite misleading for the political power analyst (Baldwin, 1989).
Some scholars have suggested that the fungibility of power resources increases as the amount increases (Art, 1996; Waltz, 2000). Thus, power is said to be more fungible for powerful states than for weaker states. It is not clear what this means or why it might be true. It is, of course, true that more power resources allow one to do more things, that is, influence more actors and/or more issues. This implies nothing about the fungibility of any particular power resource. Fungibility refers to the uses of a given amount of a power resource, not to the uses of varying amounts. In the economic realm, rich people can buy more things than poor people; but this is not because a rich person’s dollar is more fungible than a poor person’s dollar. The contention that fungibility increases with the amount of power resources is based either on a confused concept of fungibility or on a logic that has yet to be spelled out (Baldwin, 1999; Guzzini, 1998).
The Problem of Intentions
Max Weber (1947: 152) defined power as ‘the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance, regardless of the basis on which this probability rests.’ This definition clearly makes the intentions of actor A an important part of the concept of power. Many of the most interesting and important questions in international relations concern the ability or inability of governments to realize their goals. Can the Allies win the Second World War? Can the United States get other countries to join the United Nations? Can Japan get the members of the United Nations to let it join? Can China get the approval of member countries to join the World Trade Organization? Can the poor countries get trade preferences from the rich? All such questions involve the ability of countries to realize their goals.
But what about unintended effects? When the United States Federal Reserve system raises interest rates, it usually intends to affect the American domestic economy; but the actual effects are likely to reverberate around the world. There is no question about the reality or importance of unintended effects in international politics (Guzzini, 2000; Jervis, 1997; Strange, 1988). The question is whether the conventional concept of power can account for such phenomena. Although intentions are often built into the causal concept of power, for example, the Weberian version, they need not be. It is quite possible to differentiate between situations in which A intentionally causes a change in B’s behavior and situations in which A does so unintentionally (Baldwin, 1989; Frey, 1989). Relational power analysis is historically indebted to the Weberian formulation, but it is not logically bound by it. Thus, there is no need for a fundamental reformulation of the concept of power in order to account for its unintended effects.
Those who call for more attention to the unintended effects of power tend to imply that these unintended effects are detrimental to the interests of those affected (Guzzini, 2000; Strange, 1988). This is not necessarily so. The unintended effects can also be beneficial to the interests of those affected. When the United States encourages trade with other countries, it does so primarily with the intention of improving its own economic welfare; but this may have the unintended effect of improving the welfare of its trading partners also. When the United States took steps to deter Soviet nuclear attack on North America during the Cold War, it did so primarily with the intention of providing for its own security; but this action had the unintended effect of providing for Canadian security also. Whether the unintended effects of the actions (or inactions) of powerful states tends to be beneficial or detrimental to the interests of those affected is an empirical question. It should be answered by research, not by assertion.
The Measurement Problem
Before one can measure power, one must first have a concept of power. In the field of international relations, the desire to measure power on a single dimension that would allow states to be ranked often gets in the way of–or even precedes–conceptual analysis. Frey (1989) has pointed out that the difficulty of measuring power often leads researchers to redefine it so as to make operationalization easier. ‘In this fashion, power has frequently been defined in terms of supposed resources–e.g., the ability to mobilize resources, possession of resources, and other forms of what Elster (1976: 252) calls “generalized fetichist theories,” that is, theories that attempt to regard relations as properties’ (Frey, 1989: 7–8).
As noted above, there is no political counterpart for money. There is no standardized measure that facilitates reducing the various dimensions of power to a single dimension. Yet the desire to measure power makes this an inconvenient fact:
The search for an index of national power has been largely, … based on the assumption that it is possible and desirable to find a currency of politics. As economists view economic transactions of all sorts and at all levels in terms of a standardized unit of currency, … so, the assumption runs, must the political scientist find an absolute scale along which to evaluate the ‘power’ of nation-states. (Merritt and Zinnes, 1988: 142)
It is the desire of international relations scholars to rank the overall power of states from highest to lowest that generates the most difficult measurement problems. This requires comparing different dimensions of power relations without any agreed-upon way to do this. Some scholars contend that the question of ‘Who’s number one?’ is as useful in international relations as it is in sports (Ray and Vural, 1986). It is not clear, however, that it is either meaningful or useful to ask this question even in the realm of sports. Assessing athletic ability without reference to a specified set of athletic activities is akin to assessing power without reference to scope and domain. How is one to compare a golfer, a swimmer, an archer, a runner and a weightlifter? As Dahl ([1963, 1984] 1991: 27) has pointed out, ‘it is difficult enough to estimate relative influence within a particular scope and domain; it is by no means clear how we can “add up” influence over many scopes and domains in order to arrive at total, or aggregate, influence.’ This is equally true of attempts to ‘add up’ and compare athletic accomplishments in different sports.
Most indices of overall national power rely primarily on GNP, but are sometimes supplemented with demographic and military measures (Merritt and Zinnes, 1988). The best known of these is that developed by the Correlates of War Project (Singer, 1988). The difficulty with all such measures, however, is that they treat power as a property rather than a relation. ‘The escape through redefining power to be a property, though seductive, warps the very essence of what interests us’ (Frey, 1985: 12). GNP is correlated not only with war-winning ability, but also with Olympic medals. No one, however, has suggested that GNP is a useful measure of athletic ability. Not every correlate of power (or Olympic medals) provides a useful operational definition of power (or athletic ability).
This is not to suggest that it is impossible to measure power. Power is difficult but not impossible to measure. Measures of the power of A with respect to B (domain) and with respect to C (scope) can be made on the following dimensions: (1) the probability of B’s compliance; (2) the speed with which B complies; (3) the number of issues included in C; (4) the magnitude of the positive or negative sanction provided by A; (5) the costs to A; (6) the costs to B; and (7) the number of options available to B (Dahl, 1968; Frey, 1985, 1989). If international relations researchers were to give up the search for a universally valid measure of overall national power, much useful research could be focused on measuring the distribution of power within specified scopes and domains. What makes the Correlates of War power index more useful than most such indices is that it was developed, and has usually been applied, in the context of the policy-contingency framework of war. Whether this index is equally applicable to other situations has yet to be established.
Power in International Relations Theory
‘The proposition that the nature of international politics is shaped by power relations’ is often listed as a ‘defining characteristic of Realism’ (Wendt, 1999: 96–7). As Wendt (1999: 97) points out, however, this is not a unique characteristic of realism. Neoliberals, Marxists, postmodernists, constructivists, dependency theorists, globalists and feminists all think power matters. No attempt will be made here to survey the treatments of power relations in all of these theories. The discussion will confine itself to two well-known and influential theories–the balance of power and neorealism.
Classic Balance of Power Theory
The ‘balance of power’ was used by Thucydides to explain the onset of the Peloponnesian War, was the subject of an essay by David Hume (1742) in the eighteenth century, and continues to fascinate international relations theorists even today (Claude, 1989; Guzzini, 2000; Little, 1989; Moul, 1989; Walt, 1987; Waltz, 1979). Although many different theories carry the ‘balance of power’ label, the term itself, ‘implies that changes in relative political power can be observed and measured’ (Wright, 1965: 743).
The question of precisely what was being observed and measured, however, has remained illusive. In the ninettenth century Richard Cobden argued that the term ‘balance of power’ could ‘be discarded as fallacious, since it gives no definition–whether by breadth of territory, number of inhabitants, or extent of wealth–according to which, in balancing the respective powers, each state shall be estimated’ (quoted in Gulick, 1955: 27). Pollard (1923: 58) concluded that the term ‘may mean almost anything; and it is used not only in different senses by different people, or in different senses by the same people at different times, but in different senses by the same person at the same time.’ Morgenthau (1960: 167) discussed the balance of power at length, but admitted to using the term to mean four different things. One is tempted to despair when one writer dismisses the term as meaningless (Guzzini, 2000), while another contends that the problem is ‘not that it has no meaning, but that it has too many meanings’ (Claude, 1962: 13; Haas, 1953). It is beyond the limits of this chapter to attempt clarification of this conceptual morass.
No matter which version of balance of power theory one considers, the idea of power as a property rather than a relation is firmly embedded. It could hardly be otherwise, since any attempt to interpret balance of power theory using the relational concept of power would immediately encounter the difficulties flowing from the multidimensionality of power and the lack of a standardized measure of value in terms of which these dimensions could be expressed. Suppose a country drains resources from its domestic economy in order to increase its military strength, as the Soviet Union did. Its military power may be increasing at the same time, and partly because, its economic power is decreasing. How is one to calculate the net effect on the overall balance of power, given the difficulty of adding up various scopes and domains of power? It is precisely these difficulties that lead Guzzini (1998, 2000) to pronounce the term meaningless.
To the extent that balance of power theory has been meaningful, it has been based on a conception of power as a particular type of power resource used in a particular policy-contingency framework, that is, military force conceived in the context of war-winning ability (Claude, 1962; Gulick, 1955; Morgenthau,  1960; Walt, 1987; Wright, 1965: 743ff). The analytical perspective of relational power prompts one to ask, ‘Power to get whom to do what?’ One of the benefits of bringing this perspective to bear on balance of power theories is that it brings to light the underlying assumptions that: (1) military force is the measure of power; and (2) war-winning is what matters most. Only after these assumptions have been made explicit can fruitful debate as to their wisdom occur.
The theory of neorealism developed by Waltz (1979) dominated discussions of international relations theory during the last quarter of the twentieth century, much as Morgenthau’s (1948) version of the theory of realism dominated discussions during the period between 1950 and 1975. Overall evaluation of neorealism is beyond the scope of this chapter. Instead, the focus is on the role of power and capabilities in the theory.
Waltz advances a structural theory of international politics. One of the defining characteristics of the structure of the international system is the distribution of capabilities. Since judgments must be made about how capabilities are distributed, Waltz must confront the issue of how to measure them. Realizing that his theory requires the rank ordering of states according to their capabilities, he resists the specification of scope and domain necessitated by a relational notion of power. Ranking the capabilities of states is much harder if power/capability is conceived as multidimensional. Thus, he asserts that ‘the economic, military, and other capabilities of nations cannot be sectored and separately weighed’ (1979: 131). He provides neither argument nor evidence to support the assertion that different kinds of capabilities cannot be measured separately; he simply asserts it. It may be that Waltz has in mind the constraints of his theory in the sense that permitting capabilities to be weighed separately could make ranking states excessively difficult. Waltz goes on to say that ‘states are not placed in the top rank because they excel in one way or another. Their rank depends on how they score on all of the following items: size of population and territory, resource endowment, economic capability, military strength, political stability and competence. States spend a lot of time estimating one another’s capabilities, especially their abilities to do harm’ (1979: 131). The use of the term ‘score’ is revealing. It implies a measuring rod, or standard, in terms of which the various elements of national power can be evaluated; but there is no indication of what this standard is. The assertion that states devote ‘a lot of time to estimating one another’s capabilities’ is unsupported and contestable. The defense ministries of states formulate contingency plans with respect to a variety of policy-contingency frameworks, but it is unlikely that they spend much time estimating each other’s capabilities in general or without reference to actual or postulated situations. The idea that American policy-makers spend a lot of time calculating the capabilities of Canada or the United Kingdom in general, or in the abstract, seems rather far-fetched. Still, these are empirical questions and are, in principle, researchable.
Despite his admission that ‘states have different combinations of capabilities which are difficult to measure and compare’ (1979: 131), Waltz proclaims that ‘ranking states … does not require predicting their success in war or in other endeavors. We need only rank them roughly by capability.’ This assertion, of course, begs the question of how ‘capabilities’ are to be defined–a definition that Waltz never provides. We are told only that capabilities are ‘attributes of units’ (1979: 98). Clearly, the relational concept of power or capabilities is ruled out, since that concept of power depicts capabilities as potential relationships rather than as properties of a single state (or unit). The question of ‘Capability to get whom to do what?’ is simply begged; and the power as resources concept underlying Waltz’s theory becomes apparent.
At some level, however, most international relations theorists recognize the wisdom of the Sprouts’s contention that ‘without some set of given undertakings (strategies, policies), actual or postulated, with reference to some frame of operational contingencies, actual or postulated, there can be no estimation of political capabilities’ (1965: 215). In most treatments of the elements of national power in international politics an implicit set of policy-contingency assumptions can be identified, usually having to do with military power. Just as Morgenthau’s discussion of the elements of national power implies that war-winning is the standard of judgment (Baldwin, 1993: 17–18), careful reading of Waltz generates a strong suspicion that war-winning ability is the unstated standard by which states are being ranked. Morgenthau’s contention that ‘nations active in international politics are continuously preparing for, actively involved in, or recovering from organized violence in the form of war’ ( 1960: 38) is remarkably similar to the outlook in Waltz’s Theory of International Politics. ‘The possibility that force will be used by one or another of the parties looms always as a threat in the background. In politics force is said to be the ultima ratio. In international politics force serves, not only as the ultima ratio, but indeed as the first and constant one’ (Waltz, 1979: 113). ‘The daily presence of force and recurrent reliance on it mark the affairs of nations. Since Thucydides in Greece and Kautilya in India, the use of force and the possibility of controlling it have been the preoccupations of international-political studies’ (Waltz, 1979: 186). Given the absence of any explicit standard for ‘scoring’ the capabilities of states in Waltz’s text, there is more than a little reason to suspect that war-winning is the implicit standard being applied.
Although the book is nearly devoid of references to the scholarly literature on relational power, at the end of Theory of International Politics (1979: 191–2), almost as an afterthought, Waltz launches a confusing and confused attack on the relational concept of power: ‘We are misled by the pragmatically formed and technologically influenced American definition of power–a definition that equates power with control. Power is then measured by the ability to get people to do what one wants them to do when otherwise they would not do it.’ This is a puzzling and misleading criticism. It is unclear why Waltz uses the phrases ‘pragmatically formed,’ ‘technologically influenced,’ or ‘American.’ The relational concept of power was developed by non-Americans as well as Americans (Barry, 1976; Weber, 1947) and has no intrinsically ethnocentric biases. And neither the meaning nor the significance of pragmatism and technology is self-evident or explained.
Waltz goes on to assert that ‘the common relational definition of power omits consideration of how acts and relations are affected by the structure of action,’ which is not necessarily true, and that unintended effects are ruled out of consideration, which is true of some versions of relational power but not others–as noted above.
‘According to the common American definition of power, a failure to get one’s way is proof of weakness.’ In a sense this is true. Actors that consistently try and fail to influence other actors are unlikely to be viewed as powerful. Indeed, Waltz himself appears to believe this, since he later observes that ‘the stronger get their way–not always, but more often than the weaker’ (Waltz, 1993).
Waltz then asks: ‘What then can be substituted for the practically and logically untenable definition? I offer the old and simple notion that an agent is powerful to the extent that he affects others more than they affect him.’ There are several remarkable aspects of this proposed definition of power. First, after rejecting both causal and relational concepts of power, he proposes a definition that is both causal and relational. Second, the notion proposed is similar to those espoused by Deutsch (1953, 1963) and Frey (1985), both of whom saw themselves as contributing to the development of the relational concept of power. Third, it is inconsistent with the statement in the very next paragraph that ‘the extent of one’s power cannot be inferred from the results one may or may not get.’ And fourth, the proposed concept of power seems to have little or nothing to do with the concepts of power and capability used throughout the earlier sections of the book. If capability is defined as the potential power to affect others more than one is affected by others, it is no longer a property of a single actor.
Even the critics of neorealism credit it with having enhanced the clarity and rigor of the realist theoretical tradition (Keohane, 1986). With respect to its treatment of power and capability, however, Theory of International Politics seems to have introduced a considerable amount of confusion, and contradiction.
The study of power in international relations has generated a number of issues in addition to those discussed above. Among these are the following: the role of military force, structural power and constructivism.
Many writers have commented on the preoccupation with military force by students of international politics down through the ages (Art and Waltz,  1999; Baldwin, 1989; Osgood and Tucker, 1967; Sprout and Sprout, 1945, 1962, 1965; Waltz, 1979; Wright, 1955,  1965). Although war is an important phenomenon that international relations scholars regard as their special province, the field of international relations has paid a price for its preoccupation with military force. The importance of military force has been exaggerated; the role of non-military forms of power has been underestimated; and the field of international relations has been impoverished by its insulation from studies of power in other realms.
The privileged place of military power in the study of international politics is demonstrated and reinforced by references to the ‘centrality’ of force to international politics (Art, 1996; Baldwin, 1999); to the study of power as ‘a study of the capacity to wage war’ (Cline,  1997); to force as ‘the ultimate form of power’ (Gilpin, 1975, 1981); or to international security studies as ‘the study of the threat, use, and control of military force’ (Walt, 1991: 212). Even Keohane and Nye ( 2001: 15), who have criticized the traditional emphasis on military force, depict force as dominating other means of power.
The tendency to single out force as the ultimate measuring rod to which other forms of power should be compared is anathema to the approach advocated by Lasswell and Kaplan (1950: ix, 76, 85, 92, 94). Although they gave ‘special consideration to the role of violence,’ they repeatedly denied that power rests ‘always, or even generally, on violence’; and they maintained ‘that power may rest on various bases’; that ‘none of the forms of power is basic to all the others’; and that ‘political phenomena are only obscured by the pseudosimplification attained with any unitary conception of power as always and everywhere the same.’ Despite the vigorous efforts of Lasswell and Kaplan and the tradition of relational power analysis they spawned, the contemporary literature on international relations often exhibits the same tendencies to exaggerate the role of military power as did earlier works (Baldwin, 1995; Ray and Vural, 1986; Walt, 1991; Waltz, 1979).
The preoccupation with military force in the study of international politics has led to the neglect of non-military forms of power, such as economic statecraft (Baldwin, 1985). In addition, it has ironically limited understanding of military statecraft itself. The question of when military force should be used cannot be answered without consideration of alternative instruments of statecraft (Baldwin, 1995; 1999/2000). Thus, the neglect of non-military forms of power has hampered understanding of the conditions under which military force should be used.
Structural vs. Relational Power
The relational power approach has been criticized both for neglecting the study of structural power and for its alleged inability to take account of structural power (Guzzini, 1993, 2000; Strange, 1988). To the extent that structural power is viewed as unrelated to human agency or based on a non-causal notion of power, it would be fair to say that relational power and structural power represent fundamentally different approaches to the study of power. Otherwise, the relational concept of power is quite capable of taking account of power structures.
If structural power refers to unintentional power or to power with respect to the creation and/or control of structures (Guzzini, 1993; Krasner, 1985; Strange, 1988), there is no need to seek an alternative to the relational concept of power. The first meaning can be taken care of by excluding intentionality from the concept of power, as noted above. And the second meaning of structural power can easily be accounted for by proper specification of scope and domain. The creation and/or control of structures is simply an instance of influence with a particular scope and domain.
The study of power structures does present difficulties for the relational notion of power if such structures are depicted as unidimensional and monolithic and unspecified as to scope and domain. Thus, the idea of a single power structure dominating all issue areas and all domains to an equal degree is difficult to reconcile with the relational power approach. Some discussions of ‘hegemony’ in international relations seem to imply this view. There is no reason, however, why structures, defined as persistent patterns of power relationships in specified scopes and domains, cannot be usefully studied using the relational concept of power (Frey, 1971). It is worth noting that Lasswell and Kaplan (1950) devoted a whole chapter to ‘structures.’
Constructivism vs. Rationalism
How does the debate between constructivism and rationalism intersect with power analysis in the study of international relations? It depends on which of the many versions of constructivism one examines. If constructivism is viewed as rejecting human agency and causal concepts and theories, there is very little overlap. The postmodernist followers of Michel Foucault, for example, may find the relational power approach of little interest. Subscribers to Wendt’s (1999) version of constructivism, however, will find much grist for their mill in the relational power literature. Wendt (1999: 97) divides international relations theories into those that emphasize ‘brute material forces’ as bases of power and those that view power as ‘constituted primarily by ideas and cultural contexts.’
From its inception, the relational power approach has included both material and non-material bases of power. Lasswell and Kaplan (1950: 87) cited respect, rectitude, affection and enlightenment as base values of power and influence; and they devoted a whole chapter to ‘symbols.’ And Dahl ([1963, 1984] 1991: 35) includes information, friendship, social standing and the right to make laws in addition to threats of force and money in a list of political power resources.
In addition, norms, values, ideas and cultural contexts have figured prominently in the relational power approach. Among the factors that a power analyst might want to examine in explaining power relations, Dahl (1968: 412) included values, attitudes, expectations, decision-making rules, structures and constitutions. No constructivist is more emphatic about the importance of cultural context in power analysis than are Lasswell and Kaplan (1950: 85, 94):
In particular, it is of crucial importance to recognize that power may rest on various bases, differing not only from culture to culture, but also within a culture from one power structure to another.
None of the forms of power is basic to all the others. As patterns of valuation in a culture are modified, and changes come about in the social order and technology, now one form of power and now another, plays a fundamental role. Political analysis must be contextual, and take account of the power practices actually manifested in the concrete political situation.
In sum, far from being a battleground for the dueling forces of constructivism and rationalism, power analysis may be a point of convergence for at least some members of each camp.
Power Analysis and Policy Relevance
The two dominant traditions in power analysis in international relations have been described above in terms of the elements of national power approach, which depicts power as resources, and the relational power approach, which depicts power as an actual or potential relationship. Which is more likely to be useful to policy-makers? Nye (1990: 26) suggests that the relational power approach is likely to seem ‘too ephemeral’ to ‘practical politicians and leaders.’ The idea of power as the ‘possession of resources,’ he contends, holds more appeal for policy-makers because it ‘makes power appear more concrete, measurable, and predictable’ than does the relational definition. ‘Power in this sense,’ he notes, ‘means holding the high cards in the international poker game.’
A case can be made, however, for the opposite conclusion. It is the elements of national power approach that has proved useful in the Correlates of War Project. Various studies based on this project of numerous wars during the past 500 years (Small and Singer, 1982; Stam, 1996; Wang and Ray, 1994) have produced useful knowledge about the causes and outcomes of war. Policy-makers, however, tend to have notoriously short time horizons. If they are considering going to war, it is not very helpful to point out that if they fight fifty wars during the next century, they are likely to win most of them. Nor are they likely to care much about what factors were important in most of the wars for the past 500 years. Most policy-makers are likely to be involved in only one war. They want to know whether their country is likely to win a particular war, fought in a particular context, during a particular time period, against a particular adversary. The gross inventory of American elements of national power was not only of little help in predicting the outcome of the Vietnam War, it was quite misleading. The United States may have been the greatest power in the history of the world, but it was ill-equipped to fight a guerilla war in a faraway land with language, culture and history that it understood poorly. In that situation, a relational power approach, setting the capability estimate in the context of a relevant policy-contingency framework, would probably have been more useful to American foreign policy-makers. Context matters, and policy-makers, as practical people, are likely to understand this more readily than academics. It is correct to depict the elements of power as holding the high cards in the international poker game, but it is incorrect to imply that there is only one kind of game in international politics. If the name of the game is bridge, the person with the good poker hand may be in big trouble. Policy-makers need to know the name of the game in order to evaluate the strength of their hands.
Future Research Directions
Power analysis intersects with almost every major research program in international relations. It would be impossible to identify all of the promising avenues of research for the power analyst during the next ten years or so. Those discussed here do not begin to exhaust the possibilities for fruitful research.
Power Relations as Dependent Variables
Power may be treated as either a dependent or an independent variable (Dahl, 1968). Dahl’s (1961) classic study of community power was entitled Who Governs? In this study, power was treated as a dependent variable. The study began, as the title implies, with the assumption that power was being exercised by those who govern and proceeded to ask, ‘By whom?’; ‘On what issues?’; ‘How?’; and so on. International relations scholars need to devote more attention to power as a dependent variable. Instead of focusing on how a given power distribution affects regime formation or war initiation, international relations scholars need to devote more attention to questions like ‘Who has power with respect to which other actors, on which issues?’ ‘By what means is this power exercised?’ And ‘What resources allow states to exercise this power?’ A good example of this kind of research is Cox and Jacobson’s (1973) study of influence in international organizations. They focus on the distribution of influence, different issue areas, and different time periods. They also examine the bases of power of various actors. Students of international relations need to devote more attention to treating power as a dependent variable and less to treating it as an independent variable (cf. Caporaso and Haggard, 1989).
Forms of Power
Preoccupation with military power has led students of international relations to neglect other forms of power.
Soft power The term ‘soft power’ was introduced by Nye (1990). He used it to call attention to the ability to get ‘others to want what you want’ (Nye, 1990: 31–2). Noting that this ability to affect the preferences of others ‘tends to be associated with intangible power resources such as culture, ideology, and institutions,’ he distinguished it from ‘the hard command power usually associated with tangible resources like military and economic strength.’ Insofar as this distinction calls attention to the need for students of international relations to direct attention to forms of power other than traditional military force, it has performed a useful service. The concept, however, needs clarification. Nye’s discussion confuses power resources with scope. At times, the ‘tangibility’ of power resources seems to be the defining characteristic; while at other times, the use of those resources to control agendas or preferences seems to define soft power. It is also not clear what ‘tangibility’ means. Is a threat of military force tangible? Overall, soft power is a huge conceptual misstep in the right direction.
Further research would also be helped by recognition that there is little new in the idea of soft power from the standpoint of the literature on relational power. All of the forms of soft power discussed by Nye are familiar to relational power analysts. Further research on soft power should be more firmly rooted in that literature.
Positive sanctions Positive sanctions are actual or promised rewards. Most of the research on power in international relations focuses on negative sanctions, i.e., actual or threatened punishments (Baldwin, 1971a). Despite a number of recent works on the role of positive sanctions (Cortwright, 1997; Crumm, 1995; J. Davis, 2000; P. Davis, 1999; Long, 1996; Newnham, 2000), the opportunities for further research are enormous.
Comparative influence techniques The instruments of statecraft–diplomatic, economic, military and symbolic–tend to be studied separately. This is a hindrance from the standpoint of both theory and policy relevance. Without comparative research on techniques of statecraft, theorists can say little about the utility of various policy instruments. If the success rate of economic sanctions is estimated at 34 per cent, should one conclude that policy-makers are fools for using an instrument with such a low rate of success? Or is this about the best that can be expected of any instrument of statecraft? There is little or no reliable data on comparative success rates of instruments of statecraft.
Policy-makers have little use for research findings regarding one technique of statecraft. Policymakers need information that will help them choose among alternative policy options. Thus, what they want to know is: How successful is a given policy instrument likely to be, with respect to which goals and targets, at what cost, and in comparison with which policy alternatives? Without comparative studies of techniques of statecraft, it is hard to answer such questions (Baldwin, 1999/2000).
Military force Despite the emphasis on military force in the literature on international politics, much work remains to be done. Three problems are especially deserving of further research. First, the question of whether the utility of military force is declining needs attention. The groundwork for this research was provided by Knorr (1966: 5) long ago. The basic questions to be asked were identified as follows: ‘How much has it [i.e., force] lost in utility, if there has been any loss at all? And utility for what purpose? And to whom? And under what, if not all, circumstances? And military power in all its forms and modes of employment, or only in some?’ Utility for the economist Knorr, naturally, was a function of both costs and benefits. Recent studies that purport to say something about the utility of military power while devoting little or no attention to the costs of using force can be quite misleading (e.g., Art, 1996; Art and Waltz,  1999; Pape, 1996). Second, the fungibility of military force needs further study. To what extent can military force be used to exercise influence in which situations? Although it is usually assumed that force is quite fungible with respect to military issues and conflicts, this assumption needs to be questioned. Wars and militarized conflicts come in a variety of sizes and shapes–guerilla war, civil war, limited conventional war, limited nuclear war, chemical and biological warfare, large scale nuclear warfare, deterrent situations, etc. It is not clear that the military power resources useful in one type of war can easily be transferred to another type. Thus, more studies of the use of particular types of military power in different policy-contingency frameworks are needed.
The third problem concerns the question of how to define and measure military success (Baldwin, 1999/2000). Despite the voluminous literature on war, very little attention has been devoted to explicating the concept of success. The idea that ‘every war has a winner’ is deeply embedded in the literature on military force. The persistence of the zero-sum concept of military conflict is troublesome since it is incompatible with many of the topics dominating the scholarly research agenda during the past fifty years. As Schelling (1984: 269) notes: Deterrence … is meaningless in a zero-sum context. So is surrender; so are most limited-war strategies; and so are notions like accidental war, escalation, preemptive war, and brinkmanship. And of course so are nearly all alliance relationships, arms-race phenomena, and arms control. The fact that war hurts–that not all losses of war are recoverable–makes war itself a dramatically nonzero-sum activity.
Institutions and Power
Power can be exercised in the formation and maintenance of institutions, through institutions, within and among institutions. Institutions may reflect power relations, constrain them, or provide the basis for their existence. To what extent do the World Bank and the International Monetary Fund serve as instruments of American foreign policy? To what extent does the United Nations enhance the power of some countries and constrain the power of others? To what extent does the World Trade Organization constrain US power? To what extent does it strengthen US power? How is power distributed within the European Union (Garrett and Tsebelis, 1999; Holler and Widgren, 1999; Steunenberg et al., 1999)? To what extent do international institutions exercise power rather than merely reflecting it (Mearsheimer, 1994/95)? All of these questions provide a rich research agenda for the study of institutions and power relations.
How does domestic politics affect national power? Even classic elements of national power approaches included national morale, quality of government, public support and political stability among the determinants of a country’s power (Morgenthau, 1960). Does regime type matter? Are democracies at a disadvantage in international bargaining? How, if at all, does divided government affect a country’s international bargaining position? Although the conventional realist wisdom has depicted democracy as hampering the efficient conduct of foreign policy, recent studies have called this view into question and opened new lines of research on the relationship between domestic politics and the exercise of international power (Fearon, 1994, 1998; Lake, 1992; Mansfield et al., 2000; Martin, 2000; Milner, 1997, 1998; Milner and Rosendorff, 1996; Mo, 1995).
Strategic Interaction and Bargaining
The bare-bones specification of power in terms of A causing a change in B’s behavior is compatible with strategic interaction, but it neither calls attention to strategic interaction nor requires taking it into account. This is unfortunate, since most of what interests students of international politics involves strategic interaction. One of the most important research needs is linking the relational power literature with research on international strategic interaction (e.g., Martin, 2000; Milner, 1997, 1998; Mo, 1995).
This is not to suggest, however, that game theory is the only way to analyze strategic interaction. The work of Jervis (1997), Lake and Powell (1999), Larson (1998), Schelling (1984) and others has demonstrated the value of non-mathematical approaches to strategic interaction. Game theory is a useful tool for analyzing strategic interaction, but the analysis of international strategic interaction is too important to be left to game theorists alone. As Lake and Powell observe: ‘The strategic-choice approach is theoretically inclusive … [It] provides a foundation for integrating and synthesizing many otherwise competing theories of international relations’ (1999: 6).
Distribution of Power
The question of how power is distributed needs to be studied using the relational power approach. The work of Frey (1971, 1985, 1989) is especially relevant to this line of research. Rather than striving to produce yet another global ranking of the so-called ‘overall power’ of every country in the world, scholars need to focus on power distributions within specified issue-areas and perhaps within specified regions. To the extent that persistent patterns are found, issue-relevant structures of power may be identified. Rather than trying to identify a single overall international power structure, scholars should strive to identify multiple structures of power in different issue areas. Admittedly, such research will not try to provide answers to the question of ‘Who’s number one in the game of international poker?’ But simply redirecting attention away from that kind of question would, in itself, constitute progress in international power analysis.
Power has figured importantly in discussions of international interaction since the time of Thucydides. Despite the long tradition of power analysis in international politics, scholarly agreement on the nature of power and its role in international relations is lacking. The two principal approaches to power analysis in international interaction have been the ‘power as resources’ (or ‘elements of national power’) approach and the ‘relational power’ approach. The latter was developed during the last half of the twentieth century by scholars in philosophy and a variety of social science disciplines. Both approaches are evident in contemporary international relations scholarship.
Although power is an ancient focus in the study of international relations, there are many opportunities for further research. These include (1) the treatment of power as a dependent variable; (2) the forms of power; (3) institutions and power; (4) domestic politics and power; (5) strategic interaction; and (6) power distributions in different issue areas.
Although scholarly agreement on the nature and role of power in international interaction is unlikely in the near future, research along the lines suggested above may nevertheless enhance understanding of important dimensions of international behavior.