Hostname: page-component-cd9895bd7-8ctnn Total loading time: 0 Render date: 2024-12-28T17:14:13.806Z Has data issue: false hasContentIssue false
  • English
  • Français

The limits of hegemonic stability theory

Published online by Cambridge University Press:  22 May 2009

Duncan Snidal
Duncan Snidal
Affiliation:
Assistant Professor of Political Science at the University of Chicago.
Get access

Abstract

Hegemonic stability theory has been advanced as an explanation of successful cooperation in the international system. The basis of this “hegemonic cooperation” is the leadership of the hegemonic state; its appeal rests on attractive implications about distribution. However, two distinct strands of the theory (“coercive” and “benevolent”) must be distinguished. These strands have different conceptions of hegemony and the role of hegemonic leaders and so have different implications. Both require us to assume that the underlying international issues are public goods and that the international system does not allow for collective action. The former assumption limits the theory's range of application while the likely failure of the latter means that the theory may be wrong even within this more limited range. Simple formal models demonstrate a conclusion completely at odds with hegemonic stability theory: the decline of a hegemonic power may actually lead to an outcome both collectively superior and distributively preferable than when the hegemon was at the apogee of its power. Thus hegemonic stability is, in fact, only a special case of international cooperation. Understanding cooperation in general requires less restrictive assumptions.

Type
Articles
Information
International Organization , Volume 39 , Issue 4 , Autumn 1985 , pp. 579 - 614
Copyright
Copyright © The IO Foundation 1985

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1. Proponents of hegemonic stability theory implicitly introduce it as part of a more general lament or “nostalgia” for a perceived decline of American hegemony in recent years. Viewing the postwar period of American leadership as beneficial to Western interests more generally, they associate the decline in American strength with increased disorder in the international system. The decline of 19th-century British hegemony (with World War I as the ultimate consequence) and the absence of leadership during the interwar period (which deepened the Great Depression and perhaps helped cause World War II) are invoked to show the potential dangers of such decline. Traditional concerns about the dangers of international hegemony are mitigated by a conception of nonselfish leadership (discussed below) that is built into the theory. Thus hegemonic stability theory provides a strong normative justification for maintaining that American decline is unfortunate from the perspective of all members of the international system. A separate empirical question is whether there has been a significant decline in American hegemony. For a challenge to this conventional wisdom, see Russett, Bruce, “The Mysterious Case of Vanishing Hegemony; or, Is Mark Twain Really Dead?” International Organization 39 (Spring 1985)CrossRefGoogle Scholar.

2. See Kindleberger, Charles, The World in Depression, 1929–1939 (Berkeley: University of California Press, 1974)Google Scholar, and especially Kindleberger, “Systems of International Economic Organization,” in Calleo, David, ed., Money and the Coming World Order (New York: New York University Press, 1976)Google Scholar. In the terminology of much contemporary international political economy the “regime” is the public good being provided, as discussed in Keohane, Robert, “The Demand for International Regimes,” International Organization 36 (Spring 1982)CrossRefGoogle Scholar. These institutional public goods include pressures for low tariffs, acceptance of nondiscrimination, and provision of stable monetary relations. Note that once the emphasis is shifted to the provision of regimes, issue-areas not directly involving any underlying public goods come to be treated as if derivations based on public goods assumptions apply to them. While this assumption may be fruitful for analysis, it may also be most misleading—as I discuss later.

3. Olson, Mancur, The Logic of Collective Action (Cambridge: Harvard University Press, 1965)Google Scholar.

4. The public goods interpretation is central in Kindleberger's writing. Robert Keohane, who has given the theory both its name and its prominence through his criticisms and revisions of it, also bases his use of the theory (explicitly in some places, implicitly in others) on public goods or collective action assumptions. His most relevant works include “The Theory of Hegemonic Stability and Changes in International Economic Regimes,” in Holsti, O., Siverson, R., and George, A., eds., Change in the International System (Boulder: Westview, 1980)Google Scholar; “The Demand for Regimes”; and After Hegemony: Cooperation and Discord in the World Political Economy (Princeton: Princeton University Press, 1984)Google Scholar. Although Keohane (in correspondence useful in the revision of this article) feels it misrepresents his position, I argue below that he shares fundamental assumptions with Kindleberger even as he introduces important modifications and qualifications that accord with arguments in this article. But perhaps our positions differ most pointedly in how we approach the deficiencies of the theory. Keohane's solution is to demote the theory to a nonfalsifiable “interpretive framework,” useful for description but not for explanation {After Hegemony, pp. 39, 195). He uses this framework as part of an often compelling account of postwar international cooperation, but the theory itself is left in limbo: neither right nor wrong, j ust available. I am more concerned with understanding the assumptions and hence the range of the theory, and how a revised, nontautological theory might better explain international politics.

5. Keohane (in correspondence) argues this is not a central proposition in the theory-however, to deny it is to deprive the theory of its originality and to ignore its logical basis. What is novel in the theory is not the claim that strong actors can impose regimes in international politics (which goes back at least as far as Thucydides) but the use of the collective action formulation and the implication that hegemony is more widely beneficial. Moreover, once the public goods formulation is invoked to explain the emergence of regimes under hegemony, the distributional argument follows as a logical conclusion. Indeed, the proposition (or at least the weaker form that most states benefit most of the time) seems central to Keohane's After Hegemony, which poses as a fundamental question how the benefits of hegemonic cooperation (which are generalized beyond the hegemonic actor) can be maintained after the decline of hegemony.

6. A further variation on the theory of hegemonic stability, which lies between these extremes, will be discussed below. In it, leadership is coercive but provides sufficient benefits to subordinate states that they will accept it as legitimate.

7. The normative connotation of “stability” in the everyday language of international relations can misleadingly suggest that the first proposition implies the second. To keep the propositions distinct, stability will be used here in the technical sense of an outcome that is likely to maintain itself over time. (Such stability is the result not of any inherent homeostatic property of the international system but of the configuration of states and their interests in maintaining the prevailing international equilibrium.) In this sense even a prolonged period of international conflict (e.g., the Hundred Years’ War) could be a stable outcome. The second proposition then involves a normative evaluation of the particular stable outcome. Since a stable outcome could be either good or perfectly dreadful, the two propositions are now logically distinct.

8. The correct theoretical measure would be in terms of each state's expected benefit from provision of the good. A satisfactory measure of interest is not easily available, for reasons discussed in the text, but this problem is not debilitating if it is reasonable to assume that benefits from the good are roughly proportional to country size. The exact dimension of size that is appropriate will depend on the issue-area (e.g., size of global exports and imports for trade, military power and geopolitical considerations for security issues). However, note that such measures of size are also frequently used as indicators of capability, and indeed, the hegemonic stability argument is typically transformed so that the distribution of capability substitutes for the distribution of interest. Thus a distinction not made clear in the theory (i.e., between interest and capability) is further disguised by a measurement strategy that makes no distinction. The two strands of the theory discussed later in this section are distinguished in part through the different ways in which they introduce a presumed coincidence of capability with interest. A useful discussion of related conceptual issues is McKeown, Timothy, “Hegemonic Stability Theory and Nineteenth-Century Tariff Levels in Europe,” International Organization 37 (Winter 1983)CrossRefGoogle Scholar.

9. Conybeare, John, “Tariff Protection in Developed and Developing Countries: A Cross-Sectional and Longitudinal Analysis,” International Organization 37 (Summer 1983)CrossRefGoogle Scholar; Peter Cowhey and Edward Long, “Testing Theories of Regime Change: Hegemonic Decline or Surplus Capacity?” ibid. (Spring 1983); Keohane, “Hegemonic Stability”; Lake, David, “International Economic Structure and American Foreign Economic Policy, 1887–1934,” World Politics 35 (07 1983)CrossRefGoogle Scholar; Lawson, Fred, “Hegemony and the Structure of International Trade Reassessed: A View from Arabia,” International Organization 37 (Spring 1983)CrossRefGoogle Scholar; McKeown, “Tariffs”; and Stein, Arthur, “The Hegemon's Dilemma: Great Britain, the United States, and the International Economic Order,” International Organization 38 (Spring 1984)CrossRefGoogle Scholar.

10. For a succinct discussion and demolition of Friedman's wrongheaded “irrelevance of assumptions” argument, see Blaug, Mark, The Methodology of Economics (Cambridge: Cambridge University Press, 1980), pp. 104–28Google Scholar. Even those who still cling to the irrelevance argument will surely agree that it is nevertheless important to know which assumptions are being used, what their logical implications are, and whether modifications of them provide more accurate predictions.

11. A clarification of assumptions will also help to establish the appropriate data and tests for the theory. Even though Conybeare's results in “Tariff Protection” are fairly negative, for example, the underlying hegemonic stability argument is not sufficiently closely modeled for those results to be treated as definitive.

12. See Keohane, , “Demand for Regimes,” pp. 348–50Google Scholar. For a more detailed elaboration based on the distinction between regime start-up and maintenance costs, see Lipson, Charles, “The Transformation of Trade: The Sources and Effects of Regime Changes,” International Organizatiom 36 (Spring 1982)Google Scholar. Both Lipson and Keohane show the limitations of “leadership lag” arguments.

13. Krasner, Stephen, “State Power and the Structure of International Trade,” World Politics 28 (04 1976)CrossRefGoogle Scholar.

14. Ibid., p. 335.

15. Gilpin, Robert, U.S. Power and the Multinational Corporation (New York: Basic, 1975)CrossRefGoogle Scholar.

16. Baumgartner, Tom and Burns, Tom, “The Structuring of International Economic Relations,” International Studies Quarterly 19 (06 1975)CrossRefGoogle Scholar.

17. This symbiosis between hegemony and cooperation is reflected in Keohane's assessment of the beneficiaries of postwar U.S. hegemony. Secondary states in the system have benefited greatly, probably more than the hegemonic power. Evaluating the benefits to smaller and less developed states (which are not really full members of international economic regimes) is harder, but the tentative conclusion is that they fared better than they would have done under feasible alternatives. See After Hegemony, pp. 45, 252–57.

18. Gilpin, Robert, War and Change in World Politics (Cambridge: Cambridge University Press, 1982)Google Scholar. Stein points out related similarities between A. F. K. Organski's power transition model and hegemonic stability theory in “Hegemon's Dilemma.”

19. This line of argument, evident in other writers, is most distinctive in Gilpin. Although he says that this effect is “usually weak or nonexistent” (War and Change, p. 34), he uses it as a cornerstone of his theory. Without it, as discussed below, his argument has very different implications. Otherwise, why would the hegemon worry about providing public goods rather than pursuing its own individual interests? To let this claim rest on a supposed correlation of interests between dominant and subordinate actors (as in mainstream hegemonic stability theory) is insufficient for his purposes.

20. The same problem is prevalent in democratic societies where formal institutional controls are needed so that, if the government errs in the provision of or taxation for public goods, its centralized authority can be checked from below. The unilateral direction of control under international hegemony makes it more difficult to ensure that the hegemonic state does not become exploitative.

21. Attempts to test this model have often relied on measures of relative hegemony. These measures come not from the benevolent leadership but from the coercive leadership strand of the theory. The two are sufficiently distinct that it is important to keep them straight in empirical tests. However, it is possible that the definition of collective goods intended in the benevolent model implies a good with sufficient rivalness that relative size is the relevant concept. (For a discussion of different conceptions of collective goods, see Hardin, Russell, Collective Action [Baltimore: Johns Hopkins Press, 1982], chap. 3Google Scholar.) This seems unlikely given the heavy reliance on the public goods argument but, if such is the case, a further clarification of the theory is needed.

22. Keohane, After Hegemony, McKeown, “Tariffs.”

23. Of course, it is possible that a regime will be a “public” good to a subset of states and that the theory only concerns the provision of a regime benefiting a more restricted set of states. Domestic analogies would include higher union wages as a public good to union workers but not to employers and nonunion workers, or oligopoly profits as a public good to colluding oligopolists but not to consumers. However, the claims for the virtues of hegemonic order do not follow from an analysis of such “restricted public” goods, and the implications of such a revised theory would be very different and less attractive. For different but related critiques of the centrality of “order” in the hegemonic stability and regimes literature, see Strange, Susan, ’Cave! Hic Dragones: A Critique of Regime Analysis,” International Organization 36 (Spring 1982), pp. 486–88CrossRefGoogle Scholar, and Richard Ashley, “The Poverty of Neorealism,” ibid. 38 (Spring 1984), pp. 245–48.

24. I use nonexclusion in terms of the inability to control exclusion. See Snidal, Duncan, “Public Goods, Property Rights and International Organizations,” International Studies Quarterly 23 (12 1979)CrossRefGoogle Scholar.

25. An alternative possibility is that the hegemonic actor is able to enforce exclusion over the good and makes it available only to other states that allow it to “tax” them for it. This might provide a basis for a third strand of the theory, combining properties of both benevolent and coercive models, although it is not clear that it is empirically very relevant. The various problems of achieving collectively optimal levels of provision for public goods persist in all versions of the theory.

26. Kindleberger, , “International Economic Organization,” p. 37Google Scholar. Keohane in After Hegemony parts company with Kindleberger in arguing that collective action is possible under certain circumstances.

27. This is only the briefest summary of major points in the rapidly growing literature on international cooperation. The arguments here reflect a perspective shared by such recent analyses of international cooperation as Keohane, , After Hegemony, Oye, Kenneth, ed., Cooperation under Anarchy (forthcoming as a special issue of World Politics, 10 1985)Google Scholar; and Duncan Snidal, “International Cooperation: A Game Theory Analysis of Regimes and Interdependence” (book manuscript).

28. Actually, the formal results of neoclassical theory show only that, with compensation, each state could gain from trade—although this is typically treated as implying that each will gain. Other theoretical assumptions give very different results. For example, mercantilist theories emphasize relative position and see no joint (or net) gain but only zero-sum distribution among states. Imperialism theories concede net global gains but a lop-sided distribution that makes them far from joint. Nevertheless, the neoclassical theory provides at least a widely accepted basis upon which to rest the application of hegemonic stability theory to international trade issues.

29. Indeed, it is possible to argue that exclusiveness in trade agreements may have been a fundamental building block of trade regimes along subsystemic rather than global organizing principles. For a useful discussion and historical narrative on this point, see Stein, “Hegemon's Dilemma.”

30. For a discussion of solution theory in n-person games, see Shubik, Martin, Game Theory in the Social Sciences (Cambridge: MIT Press, 1982)Google Scholar.

31. Schelling, Thomas, Micromotives and Macrobehavior (New York: Norton, 1978)Google Scholar.

32. Ibid., and Hardin, Collective Action. Hardin demonstrates how Olson's argument confuses the size of the overall group with the minimum-size group that can be effective in collective action. Schelling (Micromotives, p. 221) also has some interesting observations on how it is not always “k” but often its relation to “n” that is most important. But for the case of public goods, k alone is the most relevant criterion for successful cooperation.

33. Here I interpret size in terms of the different interests of states so that the presentation corresponds to Kindleberger's “benevolent leadership” problem. When coercive leadership is involved, size can be interpreted primarily in terms of capacity. As discussed above, hegemonic stability theory typically does not make any clear distinction between interest and capability but subsumes them both under the single notion of size. This notion is also flexible enough to accommodate differences in intensity over international issues or in efficiency in producing international public goods-although such considerations further increase the need for the relation between interest and capability to be built explicitly into the theory. Finally, the shift to unequalsized actors means that the interpretation of payoffs on the vertical axis changes from “per state” to “per capita” (within a state). Thus the total payoff to a state is equal to the product of its payoff (indicated on the vertical axis) and its size (indicated on the horizontal axis). Since we are interested primarily in whether net gains from cooperation are positive, the distinction between “per capita” and “per state” payoffs is secondary; the two will always increase or decrease together for any state.

34. The payoff curves (C and NC) may also change depending on the extent of jointness and how it is affected by the growth in the system. The analysis below uses relative size and keeps the payoff curves fixed. Thus jointness is limited: as the size of states increases uniformly, their total payoff from any given level of cooperation stays constant. The case analyzed below is, as a result, a tougher case for collective action than if benefits increased with absolute size thereby increasing the incentives for cooperation in a world of growth, ceteris paribus.

35. Changes in absolute growth can be represented by changes in the length of On and the “unit” sizes for individual states. Alternatively, if we wish to maintain relative sizes on the horizontal axis (as in Figure lc), the impact of changes in absolute size can be indicated by rotations of the C and NC payoff curves around their respective y-intercepts. Positive (average) growth rates result in counterclockwise rotation, which raises the two curves, easing the problem of cooperation by lowering the size of the minimum k-group regardless of changes in relative size. Conversely, negative absolute growth will make cooperation more difficult.

36. For example, the U.S. share of GDP for all OECD countries has dropped from 51% to 35% in the I960–80 period while the share of the largest three countries (the countries involved have changed in the period) has dropped only from 65% to 62%. The corresponding figures for export share are 25% to 18% and 51% to 45%. The combined share of the United States, West Germany, and Japan has risen slightly in both categories over this period. For a thorough analysis of whether there has been a decline, see Russett, “The Mysterious Case.”

37. The complexities of bargaining introduced by “partial” contributions toward the good are not ignored. They are reintroduced in the discussion of bargaining over the distribution of net benefits of collective action.

38. This is a modified version of the standard characteristic function. The value of each coalition is determined not by the worst conceivable outcome (as in standard characteristic functions) but by the likely outcome in an analysis of the strategic situation. For example, if J or WG does not cooperate in this example, the worst that could happen is that none of the public good is provided and each receives a zero payoff. The more likely eventuality from an analysis of the situation is that each will receive the benefit of a free ride on the contribution of US.

39. The problem of “burden sharing” in cooperation further accentuates the difficulties for two-state coalitions in this example. For either WG or J to be net beneficiaries from cooperation, the hegemonic state would have to provide a minimum subsidy (side-payment) of 9.3 against the smaller cooperator's contribution of 12. While US would still gain from such an arrangement, it undoubtedly would provide a source of friction that the hegemonic actor has to provide over three-quarters of the smaller cooperator's “contribution.”

40. The different distributions could be achieved either through side-payments or through states (relaxing the simplifying assumption used above) varying their levels of provision of the good. Without such possibilities for redistributing the gains from collective action, cooperation will not occur since the distribution that results (in the first row of Table 4) does not provide incentives for WG or J to participate.

41. Since we implicitly hold the size of the system constant in the numerical example, this change means that the hegemonic power has declined in absolute terms while the two subordinate powers have grown. The benefit table is constructed to hold the overall gains from cooperation for the group constant. These assumptions about the nature of the hegemonic decline make cooperation substantially more difficult than the sort of “decline” that the United States is alleged to have experienced by proponents of the theory.

42. A purely logical argument could be substituted for the more empirically oriented argument of this paragraph to show why two-state coalitions are unstable. If such a coalition (including US and one of WG or J) does form, then provision of the good is not impeded by the decline of hegemony since provision actually increases from 8 to 9 units. Thus hegemonic stability theory requires that these coalitions be unstable. In the next paragraph I show that if this assumption is correct, then the grand coalition is likely to form. Hegemonic stability theory is in either case contradicted by the example since the level of provision rises as hegemony wanes.

43. Consider the distribution principles proposed in Table 4. The first and third principles now both result in a (23.4, 11.7, 11.7) payoff while the second and fourth lead to a (15.6, 15.6, 15.6) payoff. While there will still be some dispute, the bargaining range is less than in the previous case and none of the states can afford not to agree on some distribution since the baseline is so undesirable.

44. Shubik, Game Theory.