¹ Statistics were taken from Maddison, Angus, Dynamic Forces in Capitalist Development: A Long-Run Comparative View (Oxford: Oxford University Press, 1991)Google Scholar; and O'Rourke, Kevin and Williamson, Jeffrey G., Globalization and History: The Evolution ofa Nineteenth-CenturyAtlantic Economy (Cambridge, Mass.: MIT Press, 1999)Google Scholar.

² See for example Lenin, V. I., Imperialism: The Highest Stage of Capitalism (1916; New York: International Publishers, 1993)Google Scholar; Waltz, Kenneth N., Theory of International Politics (New York: McGraw-Hill, 1979)Google Scholar; Buzan, Barry, “Economic Structure and International Security: The Limits of the Liberal Case,” Economic Structure and International Security: The Limits of the Liberal Case, 38 (Autumn 1984)Google Scholar; Copeland, Dale C., “Economic Interdependence and War: A Theory of Trade Expectations,” Economic Interdependence and War: A Theory of Trade Expectations, 20 (Summer 1996)Google Scholar; Ripsman, Norrin M. and , Jean-MarcBlanchard, E, “Commercial Liberalism Under Fire: Evidence from 1914 and 1936,” Security Studies 6 (Winter 1996–97)CrossRefGoogle Scholar; Rowe, David M., “World Economic Expansion and National Security in Pre-World War I Europe,” World Economic Expansion and National Security in Pre-World War I Europe, 53 (Spring 1999)Google Scholar; idem, “The Tragedy of Liberalism: How Globalization Caused the First World War,” The Tragedy of Liberalism: How Globalization Caused the First World War, 14 (Spring 2005)Google Scholar; Mearsheimer, John J., The Tragedy of Great Power Politics (New York: Norton, 2001)Google Scholar; and Jervis, Robert, “Theories of War in an Era of Leading Power Peace,” Theories of War in an Era of Leading Power Peace, 96 (March 2002)Google Scholar.

³ Edward D. Mansfield and Brian M. Pollins write, “Instead, large-N studies have focused almost exclusively on the past half century and have largely ignored whether and how the effects of interdependence have changed over time”; Mansfield, and Pollins, , “Interdependence and Conflict: An Introduction,” in Mansfield, and Pollins, , eds., Economic Interdependence and International Conflict: New Perspectives on an Enduring Debate (Ann Arbor: University of Michigan Press, 2003), 8CrossRefGoogle Scholar.

⁴ In separate studies Oneal and Russett, Russett and Oneal, and Barbieri examine samples between 1870 and 1939. However, all three studies aggregate data before and after World War I into a single sample. Oneal, John R. and Russett, Bruce, “The Kantian Peace: The Pacific Benefits of Democracy, Interdependence, and International Organizations, 1885–1992,” The Kantian Peace: The Pacific Benefits of Democracy, Interdependence, and International Organizations, 1885–1992, 52 (October 1999)Google Scholar; Russett, Bruce and Oneal, John R., Triangulating Peace: Democracy, Interdependence, and International Organizations, (New York: W. W. Norton and Company, 2001)Google Scholar; Barbieri, Katherine, The Liberal Illusion: Does Trade Promote Peace? (Ann Arbor: University of Michigan Press, 2002)CrossRefGoogle Scholar.

⁵ O'Rourke and Williamson (fn. 1).

⁶ We define the level of free trade in an economy as inversely related to the extent of government regulation of international commerce. As aggregate tariff levels rise, the quantity of imports entering the country duty free should decline.

⁷ For empirical evidence supporting this claim, see Gowa, Joanne and Mansfield, Edward D., “Alliances, Imperfect Markets, and Major Power Trade,” Alliances, Imperfect Markets, and Major Power Trade, 58 (October 2004)Google Scholar; and Keshk, Omar M. G., Pollins, Brian M., and Reuveny, Rafael, “Trade Still Follows the Flag: The Primacy of Politics in a Simultaneous Model of Interdependence and Armed Conflict,” Trade Still Follows the Flag: The Primacy of Politics in a Simultaneous Model of Interdependence and Armed Conflict, 66 (November 2004)Google Scholar.

⁸ Polachek, Solomon W., “Conflict and Trade,” Conflict and Trade, 24 (March 1980)Google Scholar; Domke, William, War and the Changing Global System (New Haven, Conn.: Yale University Press, 1988)Google Scholar; Mansfield, Edward D., Power, Trade, and War (Princeton: Princeton University Press, 1994)Google Scholar; Oneal, John R. and Russett, Bruce, “The Classical Liberals Were Right: Democracy, Interdependence, and Conflict, 1950–1985,” The Classical Liberals Were Right: Democracy, Interdependence, and Conflict, 1950–1985, 41 (June 1997)Google Scholar; Oneal and Russett (fn. 4); and Russett and Oneal (fn. 4). For recent edited volumes exploring the links between globalization and military conflict, see Mansfield and Pollins, eds. (fn 3); and Schneider, Gerald, Barbieri, Katherine, and Gleditsch, Nils-Petter, eds., Globalization and Armed Conflict (Lanham, Md.: Rowman and Littlefield, 2003)Google Scholar. For important criticisms of these claims, see Barbieri (fn. 4) and Keshk, Pollins, and Reuveny (fn. 7).

⁹ Morrow, James D., “How Could Trade Affect Conflict?” How Could Trade Affect Conflict? 36 (July 1999)Google Scholar; idem, “Assessing the Role of Trade as a Source of Costly Signals,” in Mansfield and Pollins, eds. (fn. 3); Erik Gartzke, “The Classical Liberals Were Just Lucky: A Few Thoughts About Interdependence and Peace,” in Mansfield and Pollins, eds. (fn. 3); Gartzke, Erik, Li, Quan, and Boehmer, Charles, “Investing in the Peace: Economic Interdependence and International Conflict,” Investing in the Peace: Economic Interdependence and International Conflict, 55 (Spring 2001)Google Scholar; and Reed, William, “Information and Economic Interdependence,” Information and Economic Interdependence, 47 (February 2003)Google Scholar.

¹⁰ A critical paper examining the integration of bargaining models into the study of war is Fearon, James D., “Rationalist Explanations for War,” International Organization (Summer 1995)CrossRefGoogle Scholar.

¹¹ For similar critiques, see Rowe (fn. 2, 1999, 2005); Gerald Schneider and Gunther G. Schulze, “The Domestic Roots of Commercial Liberalism: A Sector-specific Approach,” in Schneider, Barbieri, and Gleditsch (fn. 8); and McDonald, Patrick J., “Peace Through Trade or Free Tta&zi” Journal of Conflict Resolution (August 2004)CrossRefGoogle Scholar.

¹² Ricardo, David, The Principles of Political Economy and Taxation (1821; London: Dent, 1973)Google Scholar.

¹³ Flam, Harry and Flanders, M. June, Heckscher-Ohlin Trade Theory (Cambridge, Mass.: MIT Press, 1991)Google Scholar.

¹⁴ Jones, Ronald, “A Three Factor Model in Theory, Trade, and History,” in Bhagwati, Jagdish, Jones, Ronald, Mundell, Ronald A., and Vanek, Jaroslav, eds., Trade, Balance of Payments, and Growth (Amsterdam: North-Holland, 1971)Google Scholar.

¹⁵ Stein, Arthur A., “Governments, Economic Interdependence, and International Cooperation,” in Tetlock, Philip E., Husbands, Jo L., Jervis, Robert, Stern, Paul C., and Tilly, Charles, eds., Behavior, Society, and Nuclear War, vol. 3 (New York: Oxford University Press, 1993)Google Scholar; idem, “Trade and Conflict: Uncertainty, Strategic Signaling, and Interstate Disputes,” in Mansfield and Pollins, eds. (fn.3); McMillan, Susan, “Interdependence and Conflict,” Interdependence and Conflict, 41 (May 1997)Google Scholar; Barbieri, Katherine and Schneider, Gerald, “Globalization and Peace: Assessing New Directions in the Study of Trade and Conflict,” Globalization and Peace: Assessing New Directions in the Study of Trade and Conflict, 36 (July 1999)Google Scholar; Mansfield and Pollins in Mansfield and Pollins, eds. (fn. 3); and Beth Simmons, “Pax Mercatoria and the Theory of the State,” in Mansfield and Pollins, eds. (fn. 3). For exceptions to this neglect of domestic politics, see Solingen, Etel, Regional Orders at Century's Dawn: Global and Domestic Influences on Grand Strategy (Princeton: Princeton University Press, 1998)Google Scholar; Papayoanou, Paul, Power Ties: Economic Interdependence and War, (Ann Arbor: University of Michigan Press, 1999)CrossRefGoogle Scholar; Rowe (fn. 2 1999, 2005); Christopher Gelpi and Joseph M. Grieco, “Economic Interdependence, the Democratic State, and the Liberal Peace,” in Mansfield and Pollins, eds. (fn. 3); and Schneider and Schulze (fn. 11).

¹⁶ See for example Fearon (fn. 10); Powell, Robert, In the Shadow of Power: States and Strategies in International Politics (Princeton: Princeton University Press, 1999)Google Scholar; and Gartzke, Erik, “The Capitalist Peace,” American Journal of Political Science (January 2007)CrossRefGoogle Scholar.

¹⁷ Fearon (fn. 10).

¹⁸ Moravscik, Andrew, “A Liberal Theory of International Politics,” A Liberal Theory of International Politics, 51 (Autumn 1997), 544Google Scholar.

¹⁹ For a brief overview of the many gains from trade including economics of scale, expanded consumer choice, more competitive prices in domestic markets, and real income growth, see Irwin, Douglas, Free Trade under Fire (Princeton: Princeton University Press, 2002), 29–39Google Scholar.

²⁰ Levinsohn, James, “Testing the Imports-As-Market-Discipline Hypothesis,” Journal of International Economics 35 (August 1993)CrossRefGoogle Scholar.

²¹ Different predictions about the distributional effects of globalization between the Heckscher-Ohlin and Specific Factors model of trade derive largely from the different assumptions about the costs of moving a factor from one sector to another. The Heckscher-Ohlin model assumes zero costs and has been used to predict that societal cleavages created by globalization depend on class. Owners of abundant factors favor globalization, while owners of scarce factors oppose it. However, the specific factors model becomes more appropriate as these costs become positive. Under these circumstances, societal cleavages are more likely to be based on industry differences. For an example of class-based analysis, see Rogowski, Ronald, Commerce and Coalitions: How Trade Affects Domestic Political Alignments (Princeton: Princeton University Press, 1989)Google Scholar. For a test of these distinctions between class and industry, see Hiscox, Michael, “Commerce, Coalitions, and Factor Mobility: Evidence from Congressional Votes on Trade Legislation,” Commerce, Coalitions, and Factor Mobility: Evidence from Congressional Votes on Trade Legislation, 96 (September 2002)Google Scholar. We will not explore these distinctions any further here. Rather we wish to exploit the simple insight that globalization produces societal conflict and not all domestic groups will necessarily support the further integration of national markets into the global economy. Our analysis is similar to that of Solingen (fn. 15) who focuses on the domestic struggle between internationalist and statist-nationalist-confessional coalitions.

²² Again, this depends on the assumptions made about the costs of moving across industries. As costs approach zero, owners of abundant factors employed in the exporting industry will see income rise. As the costs of redeployment increase, the income of factors specific to the exporting industry will rise.

²³ We distinguish between export-oriented firms and firms that are competitive in international markets without state assistance for the following reason: Firms or industries may orient their sales to international markets but rely on state assistance to be competitive there. For reasons discussed elsewhere in this article, such firms have been captured by the state and are less likely to support restrained national interests and peace. Prussian landowners prior to World War I are an example of this dynamic. This group relied on subsidies to export grain to western Russia and supported an aggressive stance against Russia during the press war in the six months prior to July 1914.

²⁴ For evidence of the link between trading interests and foreign policy goals, see Trubowitz, Peter, Defining the National Interest: Conflict and Change in American Foreign Policy (Chicago: University of Chicago Press, 1998)Google Scholar; Solingen (fn. 15); and Fordham, Benjamin O., Building the Cold War Consensus: The Political Economy of U.S. National Security Policy, 1949–1951 (Ann Arbor: University of Michigan Press, 1998)CrossRefGoogle Scholar.

²⁵ Dykstra, David L., The Shifting Balance of Power: American-British Diplomacy in North America, 1842–1848 (Lanham, Md.: University Press of America, 1999)Google Scholar.

²⁶ Trubowitz (fn. 24), 68–75.

²⁷ LaFeber, Walter, The New Empire: An Interpretation of American Expansion 1860–1898 (Ithaca, New York: Cornell University Press, 1998), 275Google Scholar. Wheat interests lobbied President Cleveland for a peaceful resolution with Great Britain over the Venezuelan boundary dispute in 1895.

²⁸ Rock, Stephen, Why Peace Breaks Out: Great Power Rapprochement in Historical Perspective (Chapel Hill, N.C.: University of North Carolina Press, 1989), 44–45Google Scholar.

²⁹ Stolper, Wolfer F. and Samuelson, Paul A., “Protection and Real Wages,” Protection and Real Wages, 9 (November 1941)Google Scholar.

³⁰ Lenin (fn. 2); and Schumpeter, Joseph, “The Sociology of Imperialisms,” in Sweezy, Paul M., ed., and Norden, Heinz, trans., Imperialism and Social Classes (1919: New York: Augustus M. Kelley, Inc., 1951)Google Scholar.

³¹ LaFeber (fn. 27).

³² Stigler, George, “The Theory of Economic Regulation,” The Bell Journal of Economics and Management Science 2 (Spring 1971)CrossRefGoogle Scholar; and Grossman, Gene M. and Helpman, Elhanan, “Protection for Sale,” Protection for Sale, 84 (September 1994)Google Scholar.

³³ Williamson, Oliver, The Economic Institutions of Capitalism (New York: Free Press, 1975)Google Scholar.

³⁴ de Mesquita, Bruce Bueno, Morrow, James D., Siverson, Randolph M., and Smith, Alastair, “An Institutional Explanation of the Democratic Peace,” An Institutional Explanation of the Democratic Peace, 93 (December 1999)Google Scholar; and de Mesquita, Bruce Bueno, Morrow, James D., Siverson, Randolph M., The Logic of Political Survival (Cambridge, Mass.: MIT Press, 2003)Google Scholar.

³⁵ The following section draws on the arguments of Fischer and Berghahn. Fischer, Fritz, War of Illusions: German Policiesfrom1911 to 1914, trans. Jackson, Marian (New York: Norton, 1975)Google Scholar; and Berghahn, V. R., Germany and the Approach of Warin 1914, 2nd ed. (New York: St. Martin's Press, 1993)Google Scholar.

³⁶ For a good discussion of the relative breakdown of economic and political interests within Germany and England during this period, see chap. 17 in Kennedy, Paul, The Rise of the Anglo-German Antagonism, 1860–1914 (London: The Ashfield Press, 1980)Google Scholar.

³⁷ Kennedy (fn. 36); and Steiner, Zara S., Britain and the Origins of World War I (New York: St. Martins Press, 1977)CrossRefGoogle Scholar.

³⁸ Spring, D. W., “Russia and the Coming of War,” in Evans, R. J. W. and Pogge, HartmutStrandmann, von, eds., The Coming of the First World War (Oxford: Clarendon Press, 1988)Google Scholar.

³⁹ See for example Fearon (fn. 10); and Wagner, R. Harrison, “Bargaining and War,” American Journal of Political Science 44 (July 2000)CrossRefGoogle Scholar.

⁴⁰ While it is true that governments, by and large, relied heavily on tariffs for revenue during the nineteenth century, debate over public finance and the distribution of taxation generated struggles and domestic cleavages similar to those surrounding globalization. The political choice to rely on tariffs for revenue disproportionately placed the burden of public finance on the segments of society (namely the left) that spent large portions of their income on consumption, i.e., food. Thus the constituents lobbying for free trade saw the removal of tariffs as a device for making the burden of public finance more democratic. Not only would the elimination of tariffs reduce the costs of consumer goods (and thus increase the political left's real income), it would also force the state to replace this lost revenue via property and income taxes, the burdens of which would fall on wealthier segments of the electorate. For discussions of these domestic dynamics and their implications for foreign policy, see Kent, Bruce, The Spoils of War: The Politics, Economics, and Diplomacy of Reparations,1919–1932 (Oxford: Clarendon Press, 1989)Google Scholar; Stevenson, David, Armaments and the Coming of War: Europe, 1904–1914 (Oxford: Clarendon Press, 1996)Google Scholar; and Herrmann, David, The Arming of Europe and the Making of the First World War (Princeton: Princeton University Press, 1996)Google Scholar.

⁴¹ For a study that tests these arguments during the second, or post-World War II era of globalization, see McDonald (fn. 11).

⁴² Blattman, Christopher, Clemens, Michael A., and Williamson, Jeffrey, “Who Protected and Why? Tariffs around the World 1870–1938” (Paper presented to the Conference on the Political Economy of Globalization, Trinity College, Dublin, August 2002)Google Scholar.

⁴³ For a recent debate exploring the role of the GATT and the WTO in promoting trade, see Rose, Andrew K., “Do We Really Know that the WTO Increases Trade?” Do We Really Know that the WTO Increases Trade? 94 (March 2004)Google Scholar; and Subramanian, Arvind and Wei, Shang-Jin, “The WTO Promotes Trade, Strongly but Unevenly,” Working Paper, no. 10024 (Cambridge, Mass.: National Bureau for Economic Research, 2003)Google Scholar.

⁴⁴ Bairoch, Paul, “European Trade Policy, 1815–1914,” Burke, Susan, trans., in Mathias, Peter and Pollard, Sidney, eds., The Cambridge Economic History of Europe, vol. 8 (Cambridge, U.K.: Cambridge University Press, 1989)Google Scholar. Referring to the differences between these two periods, O'Rourke and Williamson (fn. 1), 29, write, “all (source italics) of the commodity market integration in the Atlantic economy after the 1860s was due to the fall in transport costs between markets, and none was due to a more liberal trade policy. In contrast, most of the commodity market integration after the 1950s was (we suspect) due to more liberal trade policy.”

⁴⁵ Bairoch (fn. 44), 57.

⁴⁶ See for example Mansfield, Edward D., Milner, Helen V., and Rosendorff, B. Peter, “Free to Trade: Democracies, Autocracies, and International Trade,” Free to Trade: Democracies, Autocracies, and International Trade, 94 (June 2000)Google Scholar.

⁴⁷ Ghosn, Faten, Palmer, Glenn, and Bremer, Stuart, “The MID 3 Data Set, 1993–2001: Procedures, Coding Rules, and Description,” Conflict Management and Peace Science 21 (2004)CrossRefGoogle Scholar. A description of the COW data and numerous datasets are available from its website. Available at http://cow2.la.psu.edu/.

⁴⁸ Jones, Daniel M., Bremer, Stuart A., and Singer, J. David, “Militarized Interstate Disputes, 1816—1992: Rationale, Coding Rules, and Empirical Patterns,” Conflict Management and Peace Science 15 (1996)CrossRefGoogle Scholar.

⁴⁹ Bennett, D. Scott and Stam, Allan C., “Research Design and Estimator Choices in the Analysis of Interstate Dyads: When Decisions Matter,” Research Design and Estimator Choices in the Analysis of Interstate Dyads: When Decisions Matter, 44 (October 2000), 658Google Scholar.

⁵⁰ In a description of the MID data set, Jones, Bremer, and Singer (fn. 47), 178, write, “We.. .based our indicator of what constitutes a revisionist state on the prevailing status quo ante prior to the onset of any militarized action and recorded as revisionist the state or states that sought to overturn the status quo ante.” This explicit coding criterion raises our confidence that the definition of national interests occurs prior to the onset of a dispute. Consequently, this variable is particularly appropriate to our theoretical framework that focuses on the first stage of a war-generating process in which a domestic struggle over the costs and benefits of globalization shapes the definition of national interests over which states are willing to employ military force to defend.

⁵¹ Blattman, Clemens, and Williamson (fn. 41).

⁵² The eight deleted colonies from the Williamson data set are: Australia, Burma, Canada, Ceylon, India, Indonesia, New Zealand, and the Philippines.

⁵³ Mitchell's data allowed us to add Belgium, Bulgaria, the Netherlands, Romania, and Switzerland to the sample. These cases account for 168 cases in the baseline model sample of 1,293. Additionally, data from both Williamson and Mitchell is available for 519 of these 1,293 cases. The bivariate correlation between these two data sources is 0.9851. Brian Mitchell, International Historical Statistics (New York: Palgrave Macmillan, various years).

⁵⁴ While annual counts ranged between thirty-three and forty-four states in the international system during this period, there were on average thirty-nine. The following thirty-two states are included in our sample: Argentina, Austria-Hungary, Belgium, Brazil, Bulgaria, Chile, China, Colombia, Cuba, Denmark, Egypt, France, Germany, Great Britain, Greece, Italy, Japan, Mexico, the Netherlands, Norway, Peru, Portugal, Romania, Russia, Serbia, Spain, Sweden, Switzerland, Thailand, Turkey, the United States, and Uruguay.

⁵⁵ We utilize version 2 of the Polity IV, which converts regime scores previously classified as “interregnum” or “transition” to the twenty-one-point scale. For a discussion of these coding rules, see http://www.cidcm.umd.edu/inscr/polity/convert.htm; and Monty G. Marshall and Keith Jaggers, Polity IV Project: Political Regime Characteristics and Transitions, 1800–2002. Available at http:// www.cidcm.umd.edu/polity.

⁵⁶ Rummel, R. J., “Libertarianism and International Violence,” Libertarianism and International Violence, 27 (March 1983)Google Scholar; Lee, JamesRay, Democracy and International Conflict: An Evaluation of the Democratic Peace Proposition (Columbia, S.C.: University of South Carolina Press, 1995)Google Scholar; Benoit, Kenneth, “Democracies Really Are More Pacific (in General),” Journal of Conflict Resolution 40 (December 1996)CrossRefGoogle Scholar; and Souva, Mark and Prins, Brandon, “The Liberal Peace Revisited: The Role of Democracy, Dependence, and Development in Militarized Interstate Dispute Initiation, 1950—1999,” International Interactions 32 (April-June 2006)CrossRefGoogle Scholar.

⁵⁷ Singer, J. David, Bremer, Stuart, and Stuckey, John, “Capability Distribution, Uncertainty, and Major Power War, 1820–1965,” in Russett, Bruce, ed., Peace, War, and Numbers (Beverly Hills: Sage Publications, 1972)Google Scholar.

⁵⁸ Bennett, D. Scott and Stam, Allan C., “EUGene: A Conceptual Manual,” International Interactions 26 (2000)CrossRefGoogle Scholar.

⁵⁹ Stinnett, Douglas M., Tir, Jaroslav, Schafer, Philip, Diehl, Paul F., and Gochman, Charles, “The Correlates of War Project Direct Contiguity Data, version 3,” Conflict Management and Peace Science 19 (2002)CrossRefGoogle Scholar.

⁶⁰ While we do not include these coefficient estimates in our tables for reasons of space, this technique includes one counter and three cubic splines in each of the models. Beck, Nathaniel, Katz, Jonathan N., and Tucker, Richard, “Taking Time Seriously in Binary Time-Series-Cross-Section Analysis,” Taking Time Seriously in Binary Time-Series-Cross-Section Analysis, 42 (October 1998)Google Scholar. To counter within-panel heteroskedasticity we employ Huber-White standard errors, which are reported in parentheses below the parameter estimates.

⁶¹ Maddala, G. S., Limited Dependent and Qualitative Variables in Econometrics (New York: Cambridge University Press, 1983)CrossRefGoogle Scholar; and Keshk, Pollins, and Reuveny (fn. 7).

⁶² Gowa, Joanne, Ballots and Bullets: The Elusive Democratic Peace, (Princeton: Princeton University Press, 1999), 65Google Scholar.

⁶³ Cederman, Lars-Erik, “Back to Kant: Reinterpreting the Democratic Peace as a Macrohistorical Learning Process,” Back to Kant: Reinterpreting the Democratic Peace as a Macrohistorical Learning Process, 95 (March 2001), 22–23Google Scholar.

⁶⁴ Russett and Oneal (fn. 4), 114.

⁶⁵ Oneal and Russett (fn. 4), 24. In a series of regressions where the sample was restricted to the observations in a given year, our tests show a positive coefficient on democracy until 1905. In the years 1906 to 1914, the coefficient is negative in 1906,1907,1912, and 1914.

⁶⁶ These suggestions about the foreign-policy implications of limited political participation in nineteenth-century democracies are consistent with the arguments of the selectorate model found in Bueno de Mesquita et al. (fn. 34). They argue that the proportion of society responsible for selecting political leaders is the crucial component of democracy that shapes decisions for war and peace. The need to cultivate political support from a majority of the population leads democratic officials to be concerned about public-policy failures. They avoid entering war when the probability of victory is low. However, when entering war, they devote more resources to the effort to ensure victory than do autocratic regimes. Peace thus emerges from democratic hesitation to wage war against powerful adversaries and the unwillingness of many states to attack democracies because of the latter's tendency to contribute more resources to victory.

⁶⁷ The median value for Protect is 0.105. Its 90^th percentile value is 0.313. In these simulations we hold all other variables constant at their medians. King, Gary, Tomz, Michael, and Wittenberg, Jason, “Making the Most of Statistical Analysis: Improving Interpretation and Presentation,” Making the Most of Statistical Analysis: Improving Interpretation and Presentation, 44 (April 2000)Google Scholar.

⁶⁸ The numbers in parentheses immediately below the shift in the likelihood of conflict are a 95 percent confidence interval on that substantive result.

⁶⁹ In these 939 cases, the bivariate correlation between openness and protection is -0.155. We relied on three sources to construct this openness indicator. First, we used Oneal and Russett's (fn. 4) bilateral-trade to GDP ratios that measure interdependence between two states in a dyad. This trade series begins in 1885. To construct a monadic openness measure, we totaled up these bilateral openness measures for all potential trading partners for each state in the international system. For example, imagine there were four states in the international system in 1890: the United States, France, Great Britain, and Spain. Then imagine that the bilateral trade ratios between the United States and these three trading partners were 0.05,0.03, and 0.02. We would create a total openness score for the United States in 1890 by totaling these three scores and getting a value of 0.10. This strategy created fairly good data coverage for the period from 1885 to 1914. Second, we relied on Mitchell (fn. 53) to add pre-1885 data for nine countries to the analysis: Brazil, Denmark, France, Germany, Italy, Spain, Sweden, Great Britain, and the United States. While Mitchell has strong data coverage on imports and exports for most countries during the pre-1885 period, national income data that matches in terms of currency and price indexing (current prices) is available only for the above nine countries. Third, we relied on Maddison for GDP indices that were matched with trade data from Mitchell to construct openness indicators for the following states in the pre-1885 period: Argentina, Austria-Hungary, Belgium, Chile, Japan, the Netherlands, and Spain. Maddison, Angus, Monitoring the World Economy 1820—1992 (Paris: Development Center for Organisation for Economic Cooperation and Development, 1995)Google Scholar.

⁷⁰ Keshk, Pollins, and Reuveny (fn. 7); and Maddala (fn. 60).

⁷¹ This estimation technique is called two-stage probit least squares (2SPLS). 2SPLS estimates the coefficients very similarly to two-stage least squares. All of the right-hand variables are used to estimate predicted values of both conflict and protection levels. These predicted values are then utilized as the instrumental variables in the final estimations of the respective regressions that use MILCONFLICT and Protect as the dependent variables. We estimated the 2SPLS model using the STATA command “cd-simeq.”The program was written by Keshk. Keshk, Omar M. G., “CDSIMEQ; A Program to Implement Two-Stage Probit Least Squares,” The STATJ Journal 3 (2003Google Scholar).

⁷² Mansfield, Milner, and Rosendorff (fn. 46) find that democratic regimes trade more with each other. These findings point to the inclusion of a control for regime type in the protection equation, with the expectation that democracies have lower tariffs.

⁷³ Because large economies have large internal markets that reduce integration with the global economy, we also control for economic size with a variable measuring a country's population and expect it to be positively correlated with protection levels. Together the lagged protect variable and the population variable satisfy the exclusion restriction for this two-stage approach. The population variable carries the extra virtue of offering an operationalization of economic size that is not correlated with our key dependent variable of interest—military conflict. In bivariate regressions with conflict as the dependent variable (not shown), the coefficient on population was not significant. For a study that examines the relationship between economic openness and economic size (as measured by population), see Alesina, Alberto and Spolaore, Enrico, The Size of Nations (Cambridge, Mass.: MIT Press, 2003)Google Scholar.

⁷⁴ Much of the identification of the protection effect comes from the lagged Protect variable. Because Protect is highly autocorrelated, endogeneity bias may not be completely purged from the analysis. Future research could usefully identify other exogenous variation in trade protection.

⁷⁵ While not described in any more detail, we also checked whether an assumption of unit homogeneity across countries was valid for our sample by incorporating fixed and random effects models. Hausman tests on both of these models revealed that the probit specification without country-specific coefficients or country-specific error terms was appropriate.

⁷⁶ Blattman, Clemens, and Williamson (fn. 42).

⁷⁷ Gleditsch, Kristian, All International Politics is Local: The Diffusion of Conflict, Integration, Democratization (Ann Arbor: University of Michigan Press, 2002)CrossRefGoogle Scholar.

⁷⁸ These regressions suppress a regional dummy for North America and thus make it the reference category for the other regional variables.

⁷⁹ We obtained similar results in the single equation framework. The coefficient on Protect was still positive and significant.

⁸⁰ A single equation framework using probit produced the same primary conclusions.

⁸¹ Mansfield (fn. 8).

⁸² The purpose of this section is not to identify a new explanation for the origins of World War I. Given this conflict's status as perhaps the most heavily studied war ever, a huge volume of research identifying multiple causes for war suggests that the conflict was in many ways overdetermined. In part, the multicausal nature of this single event underscores the importance of assessing the links between free trade and peace over this entire fifty-year period of globalization with techniques, like statistical analysis, suitable to examining a large number of cases. At the same time, the centrality of this single case to realist critiques of commercial liberalism creates a need to examine it within the context of the theoretical framework presented here.

⁸³ Russett and Oneal (fn. 4).

⁸⁴ Ibid.; and Rosecrance, Richard, The Rise of the Trading State: Commerce and Conquest in the Modern World (New York: Basic Books, 1986)Google Scholar.

⁸⁵ Writing about the transition in Russian foreign policy as the Three Emperors' League expired and Russia moved closer to France in the late 188O's, Geyer notes, “Protectionism was clearly the Achilles' heel of the Russo-German Entente. . . . There was an inevitability in the mounting economic antagonism because neither Germany nor Russia could renounce protectionism.” Geyer, Dietrich, Russian Imperialism: The Interaction of Domestic and Foreign Policy, 1860—1914, trans. Little, Bruce (Leamington Spa, U.K.: Berg, 1987), 152Google Scholar.

⁸⁶ For a good summary of this press war and its implications for great power diplomacy, see Fischer (fn. 35), 370–88.

⁸⁷ Ibid., 330.

⁸⁸ Ibid., 367.

⁸⁹ McDonald, David McLaren, United Government and Foreign Policy in Russia 1900–1914 (Cambridge, Mass.: Harvard University Press, 1992)CrossRefGoogle Scholar.

⁹⁰ Ibid., 194–195.

⁹¹ Lieven, D. C. B, Russia and the Origins of the First World War (New York: St. Martin's Press, 1983), 142–43CrossRefGoogle Scholar.

⁹² Ibid., 143.

⁹³ Spring (fn. 38), 68.

⁹⁴ Fischer (fn. 35), 367.