Nations have two methods of increasing their security: building arms and forming alliances. Both methods present different political costs that must be incurred to raise security. Building arms requires shifting economic resources to the military. Forming alliances requires abandoning interests that conflict with those of the ally. Each of these strategies produces domestic opposition. A nation's response to a threat to its security must weigh the relative attractiveness of arms versus allies, both in terms of their effects on internal politics and on their external benefits. Three cases are examined in the light of this argument. The response of Austria and France to the unification of Germany in the 1860s is the central case. Theories of alliance formation based on neorealism and the offense-defense balance predict that Austria and France should have allied against the mutual threat of Prussia. This article argues that they did not form an alliance because arming separately presented lower political costs. World Wars I and II likewise are analyzed from the perspective of the argument above.

I thank Bruce Bueno de Mesquita, Scott Gartner, Paul Huth, Jack Levy, Robert Powell, and Randolph Siverson for their comments on earlier versions of this article. I retain all responsibility for the views and interpretations presented here.

¹ The basic presentation of neorealist theory can be found inWaltz,Kenneth N.,Theory of International Politics (New York:Random House,1979)Google Scholar. The terms “chain-ganging” and “buck-passing” appear inChristensen,Thomas J. andSnyder,Jack, “Chain Gangs and Passed Bucks: Predicting Alliance Patterns in Multi polarity,”International Organization44 (Spring1990), pp.137–68CrossRefGoogle Scholar.

² Christensen and Snyder, “Chain Gangs and Passed Bucks.”

³ On the offense-defense balance, seeQuester,George H.,Offense and Defense in the International System (New York:John Wiley,1977)Google Scholar; andJervis,Robert, “Cooperation Under the Security Dilemma,”World Politics30 (spring1978), pp.167–214CrossRefGoogle Scholar, especially pp. 186–210.

⁴ Benjamin A. Most and Harvey Star present a general argument that classes of foreign policies cannot be considered in isolation from their possible substitutes in their works“International Relations Theory, Foreign Policy Substitutability, and ‘Nice’ Laws,”World Politics36 (041984), pp.383–406CrossRefGoogle Scholar, andInquiry, Logic and International Politics (Columbia:University of South Carolina Press,1988), chap. 5Google Scholar. A test of the substitution hypothesis is presented inMost,Benjamin A. andSiverson,Randolph M., “Substituting Arms and Alliances, 1870–1914: An Exploration in Comparative Foreign Policy,” inHerrmann,Charles F.,Kegley,Charles W.Jr, andRosenau,James N., eds.,New Directions in the Study of Foreign Policy (Boston:Allen and Unwin,1987), pp.131–57Google Scholar.

⁵ Although this essay addresses issues in neorealist theory, the argument advanced here is not neorealist. Preferences play a central role in the argument, and neorealist theory rejects any independent role for preferences in explanations because it contends that preferences are induced by structural position. SeeWaltz,,Theory of International Politics, pp.74–77 and 127–128Google Scholar. Still, there are clearly many points of convergence between this essay and neorealist theory (e.g., security seeking states).Morrow,James D., “Social Choice and System Structure in World Politics,”World Politics41 (101988), pp.75–97CrossRefGoogle Scholar, gives a general argument that both structure and preference are essential elements of a theory of international politics. The same argument is presented inde Mesquita,Bruce Bueno andLalman,David,War and Reason (New Haven, Conn.:Yale University Press,1992)Google Scholar.

⁶The Struggle for the Mastery in Europe, 1848–1918 (New York:Oxford University Press,1954), pp.159–60 and 196Google Scholar.

⁷ See Waltz,Theory of International Politics, chaps. 6 and 8, for the neorealist position on alliances under both bipolarity and multi polarity.

⁸ SeeSnyder,Glenn H., “The Security Dilemma in Alliance Politics,”World Politics36 (061984), pp.461–95CrossRefGoogle Scholar, for a discussion of the dual problems of entrapment and abandonment in alliances.

⁹ See Jervis, “Cooperation Under the Security Dilemma” for his discussion of the security dilemma and the balance between offense and defense. AsShimshoni,Jonathan shows in “Technology, Military Advantage, and World War I: A Case for Military Entrepreneurship,”International Security15 (Winter1991), pp.187–215CrossRefGoogle Scholar, the concept of a system-level balance between offense and defense is problematic at best. Shimshoni cogently argues that the concept of a system-level offense-defense balance submerges distinctions between tactical, operational, and strategic levels of operations and the different effects military technology can have at each level and ignores the myriad of factors beyond technology that affect the outcome of battles and wars. Military technology does have an effect on the likely success of an offensive. However, that effect is not the same for all nations at the same time or for one nation against different possible opponents or even with different war plans against the same opponent. I agree wholeheartedly with Shimshoni. The affect of technology on the outcomes of wars and planning for war cannot be summarized by a simple balance between offense and defense. Nevertheless, I will accept making such distinctions for the purposes of restating and criticizing Christensen and Snyder's argument. Judging the balance between offense and defense is a difficult matter. Levy discusses the problems with determining when offense or defense is advantaged inLevy,Jack S., “The Offensive/Defensive Balance of Military Technology: A Theoretical and Historical Analysis,”International Studies Quarterly28 (061984), pp.219–38CrossRefGoogle Scholar. Further, military organization is as important as military technology in determining when the offense has an advantage. Many measures of this balance are ex post, judging an offensive advantage after offensives are proved effective. Because leaders cannot know in advance the outcome of wars, I rely on ex ante information of perceived offensive advantages. I judge that leaders believe the offensive is advantaged if they adopt offensive war plans. This definition captures Levy's concept of “The Incentive to Strike First.”

¹⁰ This conclusion does not follow from Jervis's model of the security dilemma as a prisoners' dilemma game in “Cooperation Under the Security Dilemma.” The only equilibrium of a prisoners' dilemma is mutual defection; changing the cardinal payoffs while retaining the ordinal ranking of outcomes does not change that equilibrium. One could argue that Jervis was anticipating Taylor and Axelrod's well-known application of the Folk Theorem, showing that mutual cooperation can be an equilibrium outcome of indefinitely iterated prisoners' dilemma. The likelihood that cooperation in individual rounds can be sustained as the implementation of a tit-for-tat strategy does vary with the cardinal payoffs of the game as Jervis argues. But as Axelrod and Keohane argue, iterated prisoners' dilemma may not be an appropriate model of security affairs; seeAxelrod,Robert andKeohane,Robert O., “Achieving Cooperation Under Anarchy: Strategies and Institutions,” inOye,Kenneth A., ed.,Cooperation Under Anarchy (Princeton, N.J.:Princeton University Press,1986), pp.226–54Google Scholar. Successful war may destroy the other side's ability to retaliate, undermining tit-for-tat as an equilibrium. SeeTaylor,Michael,Anarchy and Cooperation (London:John Wiley,1976)Google Scholar; andAxelrod,Robert, “The Emergence of Cooperation Among Egoists”lmerican Political Science Review75 (061981), pp.306–18CrossRefGoogle Scholar, for a proof that tit-for-tat is a Nash equilibrium of indefinitely iterated prisoners' dilemma. Axelrod and Keohane also argue that the shadow of the future may be short in security issues (seeAxelrod, andKeohane,, “Achieving Cooperation Under Anarchy,” pp.232–33)Google Scholar. A general folk theorem can be found inFudenberg,Drew andMaskin,Eric, “The Folk Theorem in Repeated Games with Discounting and with Incomplete Information,”Econometrica54 (051986), pp.533–54CrossRefGoogle Scholar.

¹¹Christensen, andSnyder,, “Chain Gangs and Passed Bucks,” pp.140–47Google Scholar.

¹² SeeMcNeill,William H.,The Pursuit of Power: Technology, Armed Force, and Society Since 1000A.D, (Chicago:University of Chicago Press,1982), chap. 7Google Scholar, for a discussion of the effect of technological change on warfare in this period. The figures for Prussian mobilization are fromHoward,Michael,The Franco-Prussian War: The German Invasion of France, 1870–1871 (London:Methuen,1961), p.60Google Scholar.

¹³Quester,,Offense and Defense in the International System, pp.77–78Google Scholar.

¹⁴ Of the two innovations, breech-loading was probably more important. Both the Union and Confederate armies used muzzle-loading rifles during the American Civil War but apparently took little advantage of the greater range of the rifled musket on the battlefield. Lack of tactical training and control led both armies to adopt essentially Napoleonic tactics on the battlefield. Breech loading allowed both higher rates of fire and the protection of a prone position. SeeGriffith,Paddy,Battle Tactics of the Civil War (New Haven, Conn.:Yale University Press,1986)Google Scholar.

¹⁵ SeeRopp,Theodore,War in the Modern World (New York:Collier,1962), pp.161–64Google Scholar, on the tactical effect of these developments.

¹⁶Showalter,Dennis E. discusses German war plans for the Seven Weeks' War inRailroads and Rifles: Soldiers, Technology, and the Unification of Germany (Hamden, Conn.:Archon Books,1976), pp.54–6Google Scholar. Also seeCraig,Gordon A.,The Battle of Königgrätz: Prussia's Victory over Austria, 1866 (Philadelphia:J. B. Lippincott,1964), pp.26–55Google Scholar, for both sides' plans for the Seven Weeks' War and their early implementation.

¹⁷ Not all French war plans were offensive. Engineers such as General Charles Auguste Frossard drew up plans that emphasized the defensive role of fortifications, the general plan for a major war since the 1830s. But none of these plans was official, and the war ministry's plans were offensive. SeeHolmes,Richard,The Road to Sedan: The French Army 1866–70 (London:Royal Historical Society,1984), pp.165–79Google Scholar.

¹⁸Altfeld,Michael F., “The Decision to Ally: A Theory and Test,”Western Political Quarterly37 (121984), pp.523–44CrossRefGoogle Scholar. Also seeMorrow,James D., “Alliances and Asymmetry: An Alternative to the Capability Aggregation Model of Alliances,”American Journal of Political Science35 (111991), pp.904–33CrossRefGoogle Scholar, for a further extension of the Altfeld model to the formation and dissolution of alliances.

¹⁹ Formally,To see this, solve Altfeld's equations (1) and (2) for the partial derivatives of utility with respect to security and set the resulting expressions equal (see Altfeld, “The Decision to Ally”).

²⁰ Alan C. Lamborn discusses the political costs, measured as internal and external risks, created by different strategies of extraction in his incisive book,The Price of Power: Risk and Foreign Policy in Britain, France, and Germany (Boston:Unwin Hyman,1991)Google Scholar. For an analysis of the domestic political costs of war preparation, seeBarnett,Michael, “High Politics Is Low Politics: The Domestic and Systemic Sources of Israeli Security Policy, 1967–1977,”World Politics42 (071990), pp.529–62CrossRefGoogle Scholar. For an analysis of those costs of alliance policy, seeBarnett,Michael andLevy,Jack S., “Domestic Sources of Alliances and Alignments: The Case of Egypt, 1962–73,”International Organization45 (Summer1991), pp.369–95CrossRefGoogle Scholar. For further discussions of this topic, seeSorokin,Gerald, “The Politics of Alliance: Security Trade-offs in Enduring Interstate Rivalries,” Ph.D. diss.,University of Michigan,1992Google Scholar.

²¹ This discussion of alliances is based on Morrow, “Alliances and Asymmetry.” A statistical model predicting when allies honour their commitments is presented inAltfeld,Michael F. andde Mesquita,Bruce Bueno, “Choosing Sides in Wars,”International Studies Quarterly23 (031979), pp.87–112CrossRefGoogle Scholar. I have not addressed the question of autonomy versus security benefits in alliances, which is discussed in Morrow, “Alliances and Asymmetry,” because this article considers only alliances between major powers, which are symmetric in the autonomy-security trade-off model of alliances.

²²Berkowitz,Bruce D., “Realignment in International Treaty Organizations,”International Studies Quarterly27 (031983), pp.77–96CrossRefGoogle Scholar.

²³ This argument parallels the predatory theory of rule; seeLevi,Margaret,Of Rule and Revenue (Berkeley:University of California Press,1988)Google Scholar.

²⁴ SeeTaylor,,The Struggle for the Mastery of Europe, pp.99–157Google Scholar, andBridge,F. R. andBullen,Roger,The Great Powers and the European States System 1815–1914 (London:Longman,1980), pp.81–102Google Scholar, for the background to the wars of German unification.

²⁵ This discussion is based onTaylor,A. J. P.,The Hapsburg Monarchy, 1809–1918 (Chicago:University of Chicago Press,1976), pp.104–8Google Scholar.

²⁶ This section on Austrian military policy is based onRothenberg,Gunther E.,The Army of Francis Joseph (West Lafayette, Ind.:Purdue University Press,1976), pp.56–89Google Scholar, andCraig,,The Battle of Königgrätz, pp.5–12Google Scholar.

²⁷Sked,Alan,The Decline and Fall of the Habsburg Empire 1815–1918 (New York:Longman,1989), p.179Google Scholar.

²⁸ SeeTaylor,,The Hapsburg Monarchy, pp.35–36Google Scholar, on the threat Italy posed to the “Austrian idea.” Also seeBarker,Nancy Nichols, “Austria, France, and the Venetian Question, 1861–66,”Journal of Modem History36 (061964), pp.145–54CrossRefGoogle Scholar;Clark,Chester Wells,Joseph,Franz andBismarck: The Diplomacy of Austria Before the War of 1866 (Cambridge, Mass.:Harvard University Press,1934), pp.40–47, 150–52, 263–66, 303–307, 403–14, and 433–37Google Scholar; andPottinger,E. Ann,Napoleon III and the German Crisis 1865–1866 (Cambridge, Mass.:Harvard University Press,1966), pp.54–9, and 104–16Google Scholar, on the difficulties of a Franco-Austrian alliance.

²⁹Decsy,János,Prime Minister Gyula Andrássy's Influence on Hapsburg Foreign Policy During the Franco-German War of 1870–1871 (New York:Eastern European Quarterly,1979), p.14Google Scholar.

³⁰Taylor,,The Hapsburg Monarchy, pp.121–40Google Scholar.

³¹ This history is based onBagdasarian,Nicholas Der,The Austro-German Rapprochement, 1870–1879 (Cranbury, N.J.:Associated University Presses,1976), pp.28–46Google Scholar;Bridge,F. R.,The Hapsburg Monarchy Among the Great Powers, 1815–1918 (New York:St. Martin's,1990), pp.72–103Google Scholar; Decsy, Prime Minister Gyula Andrássy's Influence on Hapsburg Foreign Policy During the Franco-German War of1870–1871; andTaylor,,The Hapsburg Monarchy, pp.141–8Google Scholar.

³²Plessis,Alain,The Rise and Fall of the Second Empire, 1852–1871, trans.Mandelbaum,Jonathan (Cambridge:Cambridge University Press,1985), pp.158–59Google Scholar.

³³ SeePlessis,,The Rise and Fall of the Second Empire, 1852–1871, pp.1–57 and 152–69 (election figures on p. 158)Google Scholar;Bury,J. P. T.,Napoleon III and the Second Empire (London:English Universities Press,1964)Google Scholar;Case,Lynn M.,French Opinion on War and Diplomacy During the Second Empire (Philadelphia:University of Pennsylvania Press,1954), pp.1–14Google Scholar; andZeldin,Theodore,The Political System of Napoleon III (London:Macmillan,1958), pp.80–4, 92–5Google Scholar.

³⁴ See Pottinger,Napoleon III and the German Crisis 1865–1866, andTaylor,,The Struggle for the Mastery of Europe, pp.166–75Google Scholar.

³⁵Holmes,,The Road to Sedan, p.151Google Scholar.

³⁶ This and the following paragraphs are based on Holmes,The Road to Sedan.

³⁷ These figures are fromHolmes,,The Road to Sedan, p.92Google Scholar.

³⁸Case,,French Opinion on War and Diplomacy During the Second Empire, pp.233–40Google Scholar.

³⁹ The French also produced themitrailleuse, a primitive machine gun. However, secrecy about its development and production led both to the lack of consideration of how it should be employed in combat and the lack of familiarity among the troops who operated it. It also was plagued by unreliability and other technical problems. It did not have a decisive influence on the battlefield.

⁴⁰ To parallel Christensen and Snyder, “Chain Gangs and Passed Bucks,” I discuss just the few years before the beginning of each war. Although Christensen and Snyder never state explicitly what time periods they address, their analysis covers 1911–1914 for World War I and 1938–1940 for World War II. Similarly, my analysis focuses on the immediate prewar years.

⁴¹ The chapter inThe Pursuit of Power on the relationship between arms technology and industrial production before World War I reflects the dramatic changes in naval technology compared with the lack of change in military technology for armies. SeeMcNeill,,The Pursuit of Power, pp.262–306Google Scholar.

⁴²Stone,Norman,The Eastern Front 1914–1917 (New York:Charles Scribner,1975), chap. 1Google Scholar.

⁴³ SeeKitchen,Martin,A Military History of Germany (Bloomington:Indiana University Press,1975), pp.185–87Google Scholar; andLamborn,Alan C., “Risk and Foreign Policy Choice,”International Studies Quarterly29 (121985), pp.385–410CrossRefGoogle Scholar.

⁴⁴Krumeich,Gerd,Armaments and Politics in France on the Eve of the First World War: The Introduction of Three-Year Conscription 1913–1914, trans.Conn,Stephen (Worcester, Mass.:Berg Publishers,1984)Google Scholar.

⁴⁵Rothenberg,,The Army of Francis Joseph, pp.164–65Google Scholar.

⁴⁶Ibid, pp. 157–59.

⁴⁷Taylor,,The Struggle for the Mastery of Europe, pp.487–89Google Scholar.

⁴⁸ SeeKennedy,Paul,The Rise of the Anglo-German Antagonism, 1860–1914 (London:George Allen and Unwin,1980), pp.321–60Google Scholar; andWilliamson,Samuel R.Jr,The Politics of Grand Strategy: Britain and France Prepare for War, 1904–1914 (London:Ashfield Press,1990), pp.284–307Google Scholar.

⁴⁹ SeeWatt,Donald Cameron,Too Serious a Business: European Armed Forces and the Approach to the Second World War (Berkeley:University of California Press,1975), pp.59–84Google Scholar, on the dramatic changes in military technology and doctrine in the 1930s.

⁵⁰Taylor,A. J. P.,The Origins of the Second World War (New York:Atheneum,1961), pp.246–47Google Scholar.

⁵¹ SeeMurray,Williamson,The Change in the European Balance of Power, 1938–1939: The Path to Ruin (Princeton, N.J.:Princeton University Press,1984), pp.162–66Google Scholar, for French estimates of German strength before the Sudetenland crisis.

⁵²Young,Robert J.,In Command of France: French Foreign Policy and Military Planning, 1933–1940 (Cambridge, Mass.:Harvard University Press,1978), pp.16–17CrossRefGoogle Scholar.

⁵³Hochman,Jiri,The Soviet Union and the Failure of Collective Security, 1934–1938 (Ithaca, N.Y.:Cornell University Press,1984), p.60Google Scholar. Hochman refers to Tofer as the Foreign Minister of Estonia, but the index to the third series ofDocuments on British Foreign Policy, Series 3, vol.10 (London:Her Majesty's Stationery Office,1961)Google Scholar states that Tofer was minister in Berlin from 1936 to 1939.

⁵⁴Watt,Donald Cameron,How War Came (New York:Pantheon Books,1989), p.222Google Scholar.

⁵⁵Adamthwaite,Anthony,France and the Coming of the Second World War 1936–1939 (London:Frank Cass,1977), pp.73 and 235–36Google Scholar.

⁵⁶Young,,In Command of France, p.237Google Scholar.

⁵⁷Young,,In Command of France, p.198Google Scholar.Adamthwaite,,France and the Coming of the Second World War 1936–1939, p.162Google Scholar, paraphrases this comment as fifteen days for the annihilation of the French air force.

⁵⁸ Alan S. Milward cites French defense expenditures of £215 million in 1938 and £750 million in 1939 inWar, Economy, and Society 1939–1945 (Berkeley:University of California Press,1977), p.44Google Scholar. Although these figures may be contaminated by additional expenditures after war began, clearly French military spending rose dramatically in 1939.Adamthwaite,, inFrance and the Coming of the Second World War 1936–1939, p.164Google Scholar, cites military expenditures as comprising 8.2 percent of French national income in 1938 and 22.8 percent in 1939. Adamthwaite also discusses plan 5 on pp. 162–63.

⁵⁹Adamthwaite,,France and the Coming of the Second World War 1939–1939, p.340Google Scholar. Military estimates by the French general staff were almost certainly influenced by political considerations as well as purely military factors. Still, the military was unwilling to tell the politicians that it was ready to fight in 1938 but was willing to do so in 1939.

⁶⁰ SeeHughes,Jeffrey L., “The Origins of World War II in Europe: British Deterrence and German Expansionism,” inRotberg,Robert I. andRabb,Theodore K., eds.,The Origin and Prevention of Major Wars (New York: Cambridge University Press,1988), pp.281–321Google Scholar, on Neville Chamberlain's belief in the efficacy of strategic air power and its effect on British arms policy and diplomacy.

⁶¹Murray,,The Change in the European Balance of Power, 1938–1939, pp.66–67Google Scholar.

⁶²Overy,R. J.,The Air War 1939–1945 (London:Europa Publications,1980), pp.20–25Google Scholar.

⁶³ The aims of Soviet foreign policy during the 1930s are a matter of some controversy.Posen,Barry R. identifies two views of Soviet intentions in “Competing Images of the Soviet Union,”World Politics39 (071987), pp.579–97CrossRefGoogle Scholar. He identifies the view that the Soviet Union was a “typical great power” withHaslam's,JonathanThe Soviet Union and the Struggle for Collective Security in Europe, 1933–39 (New York:St. Martin's,1984)CrossRefGoogle Scholar. Posen opposes this view with the opinion that the Soviet Union was a totalitarian, expansionist state, identifying that view with Hochman, The Soviet Union and the Failure of Collective Security, 1934–1938. My interpretation of Soviet foreign policy leans toward the former view of Soviet intentions but is not inconsistent with the latter. In either view, Stalin would wish to avoid fighting the Germans alone. War on the side of the Western Allies would be acceptable, provided they bore their share (or more) of the fighting. Either view supports the idea that Stalin needed to rebuild his economy and military before challenging Germany alone. It is not clear that any evidence could separate these two views of Soviet aims during this period. There are interpretations of either view that are consistent with Soviet policy. Given that we cannot eliminate either view, I have focused on the elements of Soviet policy that are consistent with both views.

⁶⁴Harrison,Mark,Soviet Planning in Peace and War 1938–1945(Cambridge:Cambridge University Press,1985), p.51CrossRefGoogle Scholar, cites that 85 percent of the Soviet aircraft industry and 60 percent of its armament industry were located in future war zones.

⁶⁵Watt,,How War Came, p.369Google Scholar.

⁶⁶ SeeUlam,Adam B.,Expansion and Coexistence: Soviet Foreign Policy, 1917–73,2d ed. (New York:Praeger,1974), pp.266–79, especially pp. 270–72 and 275–76Google Scholar, on the treaty discussions with the Western powers; andBullock,Alan,Hitler and Stalin: Parallel Lives (London:HarperCollins,1990), pp.667–86Google Scholar, on Soviet policy between the German occupation of Czechoslovakia and the invasion of Poland.

⁶⁷ For annual figures for Soviet arms production from 1930 to 1944, see Harrison,Soviet Planning in Peace and War 1938–1945, appendix 1.

⁶⁸Seaton,Albert,Stalin as Warlord (London:B. T. Batsford,1976), p.89Google Scholar.

⁶⁹ SeeErickson,John,The Road to Stalingrad: Stalin's War with Germany, vol.1 (London:Weidenfeld and Nicholson,1975), pp.15–24 and 37–46Google Scholar on changes in doctrine and training; pp. 32–37 and 62–5 on production of modern tanks and airplanes; and pp. 68–73 on the deployment of the Red Army.

⁷⁰ SeeHarrison,,Soviet Planning in Peace and War 1938–1945, pp.42–63Google Scholar, on Soviet economic planning for war.

⁷¹Seaton,,Stalin as Warlord, p.97Google Scholar, reports the following comment by General Georgy Zhukov according to I. K. Bagramyan, “It was Stalin's opinion that the superiority of the Wehrmacht was such that at first the Red Army would be unable to hold the frontier areas, let alone contain the sector of the enemy main attack, for the Germans had both battle experience and greater technical ability.”

⁷²Ibid, p. 93.

⁷³ Bueno de Mesquita forcefully makes the case for studying “big” wars within the study of all wars inMesquita,Bruce Bueno de, “Big Wars, Little Wars: Avoiding Selection Bias,”International Interactions, vol.16, no.2,1990, pp.159–69CrossRefGoogle Scholar. That issue of International Interactions also includes essays on other views of “big” wars. Woosang Kim and James D. Morrow provide the beginning of an explanation of “big” wars in the conclusion of their article“When Do Power Shifts Lead to War?”American Journal of Political Science36 (111992), pp.896–922CrossRefGoogle Scholar.