¹ Walzer, Just and Unjust Wars (New York: Basic Books, 1977), 42Google Scholar.

² Quoted in Roberts, Adam and Guelff, Richard, eds., Documents on the Laws of War (Oxford: Oxford University Press, 2004), 77.Google Scholar

³ Ibid., 302.

⁴ Ibid.

⁵ Ibid.

⁶ Ibid., 70, 215.

⁷ Mearsheimer, John J., “The False Promise of International Institutions,” International Security 19, no. 3 (1994–95)Google Scholar.

⁸ Hobbes, , Leviathan (New York: Penguin Books, 1985), 223.Google Scholar

⁹ Clausewitz, Carl von, On War (New York: Knopf, 1993), 83.Google Scholar

¹⁰ Downs, George W., Rocke, David M., and Barsoom, Peter N., “Is the Good News about Compliance Good News about Cooperation?” International Organization 50, no. 3 (1996)CrossRefGoogle Scholar.

¹¹ Franck, Thomas, The Power of Legitimacy among Nations (Oxford: Oxford University Press, 1990)Google Scholar. See also Koh, Harold Hongju, “Why Do Nations Obey International Law?” Yale Law Journal 106, no. 8 (1997)Google Scholar.

¹² Thomas, , The Ethics of Destruction: Norms and Force in International Relations (Ithaca, N.Y.: Cor nell University Press, 2001), 1Google Scholar.

¹³ Price, and Tannenwald, , “Norms and Deterrence: The Nuclear and Chemical Weapons Taboos,” in Katzenstein, Peter J., ed., The Culture of National Security: Norms and Identity in World Politics (New York: Columbia University Press, 1996)Google Scholar; Price, Richard M., “Reversing the Gun Sights: Transnational Civil Society Targets Land Mines,” International Organization 52, no. 3 (1998)Google Scholar.

¹⁴ Martin, Lisa L., “Credibility, Costs, and Institutions: Cooperation on Economic Sanctions,” World Politics 45 (April 1993)Google Scholar; Simmons, Beth A., “International Law and State Behavior,” American Political Science Review 94, no. 4 (2000)Google Scholar; Goldsmith, Jack L. and Posner, Eric A., The Limits of Interna-tional Law (Oxford: Oxford University Press, 2005), 100–104.Google Scholar

¹⁵ Lipson, , “Why Are Some International Agreements Informal?” International Organization 45, no. 4 (1991), 508–9Google Scholar. For a more skeptical view of the role of reputation, see Downs, George W. and Jones, Michael, “Reputation, Compliance, and International Law,” Journal of Legal Studies 31 (January 2002)Google Scholar.

¹⁶ Morrow, James D. and Jo, Hyeran, “Compliance with the Laws of War: Dataset and Coding Rules,” Conflict Management and Peace Science 23, no. 1 (2006)CrossRefGoogle Scholar.

¹⁷ Morrow, , “The Institutional Features of the Prisoners of War Treaties,” International Organization 55, no. 4 (2001), 973CrossRefGoogle Scholar.

¹⁸ Chayes, Abram and Chayes, Antonia Handler, “On Compliance,” International Organization 47, no. 2 (1993)CrossRefGoogle Scholar; Goldsmith and Posner (fn. 14), 33; Morrow (fn. 17).

¹⁹ See, for example, Fearon, James D., “Domestic Political Audiences and the Escalation of International Disputes,” American Political Science Review 88, no. 3 (1994)CrossRefGoogle Scholar; Martin (fn. 14).

²⁰ Lipson, Charles, Reliable Partners: How Democracies Have Made a Separate Peace (Princeton: Princeton University Press, 2003)Google Scholar. For the argument that democracies may make less reliable alliance partners, see Gartzke, Erik A. and Gleditsch, Kristian Skrede, “Regime Type and Commitment: Why Democracies Are Actually Less Reliable Allies,” American Journal of Political Science 48, no. 4 (2004)Google Scholar.

²¹ Schultz, Kenneth A., “Domestic Opposition and Signaling in International Crises,” American Political Science Review 92, no. 4 (1998)CrossRefGoogle Scholar.

²² Slaughter, Anne-Marie, “International Law in a World of Liberal States,” European Journal of International Law 6, no. 4 (1995)CrossRefGoogle Scholar.

²³ Russett, Bruce, Grasping the Democratic Peace (Princeton: Princeton University Press, 1993)Google Scholar; Huth, Paul K. and Allee, Todd L., The Democratic Peace and Territorial Conflict in the Twentieth Century (Cambridge: Cambridge University Press, 2002)Google Scholar.

²⁴ Poe, Steven C. and Tate, C. Neal, “Repression of Human Rights to Personal Integrity in the 1980s: A Global Analysis,” American Political Science Review 88, no. 4 (1994)Google Scholar; Davenport, Christian, “Human Rights and the Democratic Proposition,” Journal of Conflict Resolution 43, no. 1 (1999)Google Scholar; Rummel, Rudolph J., Death by Government (New Brunswick: Transaction Publishers, 1994)Google Scholar; idem, , “Democracy, Power, Genocide and Mass Murder,” Journal of Conflict Resolution 39, no. 1 (1995)Google Scholar; Harff, Barbara, “No Lessons Learned from the Holocaust? Assessing Risks of Genocide and Political Mass Murder since 1955,” American Political Science Review 97, no. 1 (2003)Google Scholar; Davenport, Christian and Armstrong, David, “Democracy and the Violation of Human Rights: A Statistical Analysis from 1976 to 1996,” American Journal of Political Science 48, no. 3 (2004)Google Scholar. Matthew Krain, however, finds no consistent relationship between regime type and the onset or severity of genocides. See Krain, , “State-Sponsored Mass Murder: The Onset and Severity of Genocides and Politicides,” Journal of Conflict Resolution 41, no. 3 (1997)Google Scholar.

²⁵ Russett (fn. 23), 33.

²⁶ See Rummel, Rudolph J., “Libertarianism and International Violence,” Journal of Conflict Resolution 27, no. 1 (1983)Google Scholar; Benoit, Kenneth, “Democracies Really Are More Pacific (in General): Reexamining Regime Type and War Involvement,” Journal of Conflict Resolution 40, no. 4 (1996)Google Scholar; Rummel, Rudolph J., “Democracies Are Less Warlike than Other Regimes,” European Journal of International Relations 1, no. 4 (1995)Google Scholar; Huth and Allee (fn. 23).

²⁷ Rummel (fn. 24, 1995).

²⁸ Reiter, Dan and Stam, Allan C. III, Democracies at War (Princeton: Princeton University Press, 2002)Google Scholar; Huth and Allee (fn. 23); Mesquita, Bruce Bueno de, Smith, Alastair, Siverson, Randolph M., and Morrow, James D., The Logic of Political Survival (Cambridge: MIT Press, 2003)Google Scholar; Siverson, Randolph M., “Democracies and War Participation: In Defense of the Institutional Constraints Argument,” European Journal of International Relations 1, no. 4 (1995)Google Scholar; Rummel (fn. 26,1995).

²⁹ Dowries, Alexander B., “Desperate Times, Desperate Measures: The Causes of Civilian Victimization in War,” International Security 30, no. 4 (2006)Google Scholar. See also Reiter and Stam (fn. 28).

³⁰ Mueller, John, War, Presidents and Public Opinion (New York: John Wiley, 1973)Google Scholar; Gartner, Scott Sigmund and Segura, Gary M., “War, Casualties, and Public Opinion,” Journal of Conflict Resolution 42, no. 3 (1998)Google Scholar; Larson, Eric, Casualties and Consensus: The Historical Role of Casualties in Domestic Support for U.S. Military Operations (Santa Monica, Calif.: RAND Corporation, 1996)Google Scholar. Feaver, Gelpi, and Reiner argue, by contrast, that casualty sensitivity in the U.S. is a function of the public's perceptions about the rightness/wrongness of the war and the success/failure of ongoing operations. See Feaver, Peter, Gelpi, Chris, and Reifler, Jason, “Success Matters: Casualty Sensitivity and the War in Iraq,” International Security 30, no. 3 (2005–6)Google Scholar.

³¹ Mesquita, Bruce Bueno de and Siverson, Randolph M., “War and the Survival of Political Leaders: A Comparative Political Analysis of Regime Types and Accountability,” American Political Science Review 89, no. 4 (1995)Google Scholar; Reiter, Dan and Stam, Allan C. III, “Democracy, War Initiation, and Victory,” American Political Science Review 92, no. 2 (1998)Google Scholar; Reiter and Stam (fn. 28). Hein Goemans, however, finds that while democratic leaders are more likely to lose power following the loss of a major war, leaders of semirepressive regimes are more likely to suffer severe punishments such as exile or execution. See Goemans, , “Fighting for Survival: The Fate of Leaders and the Duration of War,” Journal of Conflict Resolution 44, no. 5 (2000)Google Scholar; idem, , War and Punishment: The Causes of War Termination and the First World War (Princeton: Princeton University Press, 2000)Google Scholar.

³² Valentino, Benjamin, Huth, Paul, and Balch-Lindsay, Dylan, “Draining the Sea: Mass Killing and Guerrilla Warfare,” International Organization 58, no. 2 (2004)Google Scholar; Valentino, Benjamin, Final Solutions: Mass Killing and Genocide in the 20th Century (Ithaca, N.Y.: Cornell University Press, 2004)Google Scholar.

³³ Reiter, and Meek, , “Determinants of Military Strategy, 1903–1994: A Quantitative Empirical Test,” International Studies Quarterly 43, no. 2 (1999)Google Scholar.

³⁴ Downes (fn. 29), 160, 176.

³⁵ Tajfel, Henri and Turner, John, “The Social Identity Theory of Intergroup Behavior,” in Worchel, Stephen and Austin, William, eds., Psychology of Intergroup Relations (Chicago: Nelson-Hall, 1986)Google Scholar.

³⁶ Rabbie, J. M., “Group Processes as Stimulants of Aggression,” in Groebel, Jo and Hinde, Robert, eds., Aggression and War (Cambridge: Cambridge University Press, 1989)Google Scholar.

³⁷ Harmon-Jones, Eddie, Greenberg, Jeff, Solomon, Sheldon, and Simon, Linda, “The Effects of Mortality Salience on Intergroup Bias between Minimal Groups,” European Journal of Social Psychology 26, no. 4 (1996)3.0.CO;2-2>CrossRefGoogle Scholar.

³⁸ Dower, John W., War without Mercy: Race and Power in the Pacific War (New York: Pantheon, 1986), 11.Google Scholar

³⁹ Huntington, Samuel P., The Clash of Civilizations and the Remaking of the World Order (New York: Simon and Schuster, 1996)Google Scholar.

⁴⁰ Henderson, Errol A., “The Democratic Peace through the Lens of Culture, 1820–1989,” International Studies Quarterly 42, no. 3 (1998)Google Scholar; idem, , “Culture or Contiguity: Ethnic Conflict, the Similarity of States, and the Onset of War, 1820–1989,” Journal of Conflict Resolution 41, no. 5 (1997)Google Scholar.

⁴¹ Valentino, Huth, and Balch-Lindsay (fn. 32).

⁴² A large literature has emerged describing war as an ongoing bargaining process between combatants. For recent examples, see Powell, Robert, “Bargaining and Learning while Fighting,” American Journal of Political Science 48, no. 2 (2004)Google Scholar; Wagner, Harrison R., “Bargaining and War,” American Journal of Political Science 44, no. 3 (2000)Google Scholar; Slantchev, Branislav, “The Principle of Convergence in Wartime Negotiations,” American Political Science Review 97, no. 4 (2003)Google Scholar; Filson, Darren and Werner, Suzanne, “Bargaining and Fighting,” American Journal of Political Science 48, no. 2 (2004)Google Scholar.

⁴³ Schelling, , Arms and Influence (New Haven: Yale University Press, 1966), 21.Google Scholar

⁴⁴ Salisbury, Harrison E., The 900 Days: The Siege of Leningrad (New York: Da Capo Press, 1985), 516Google Scholar.

⁴⁵ Jacobs, Dan, The Brutality of Nations (New York: Paragon House, 1988)Google Scholar.

⁴⁶ Offer, Avner, The First World War: An Agrarian Interpretation (Oxford: Oxford University Press, 1989), 34.Google Scholar

⁴⁷ Home, John and Kramer, Alan, German Atrocities, 1914: A History of Denial (New Haven: Yale University Press, 2001), 74.Google Scholar

⁴⁸ Pape, Robert A., Bombing to Win: Air Power and Coercion in War (Ithaca, N.Y.: Cornell University Press, 1996)Google Scholar.

⁴⁹ Mearsheimer, , Conventional Deterrence (Ithaca, N.Y.: Cornell University Press, 1983)Google Scholar; Stam, Allan C. III, Win, Lose, or Draw (Ann Arbor: University of Michigan Press, 1996)Google Scholar; Reiter and Meek (fn. 33).

⁵⁰ Mearsheimer (fn. 49), 37.

⁵¹ See also Downes (fn. 29).

⁵² Valentino, Huth, and Balch-Lindsay (fn. 32).

⁵³ Downes (fn. 29).

⁵⁴ Liberman, Peter, Does Conquest Pay? The Exploitation of Occupied Industrial Societies (Princeton: Princeton University Press, 1996), 22–25.Google Scholar

⁵⁵ Our measure of duration, however, is only moderately correlated with our measure of military strategy at a level of .37. This relatively low level of correlation reflects the fact that some combatants attempting to implement attrition or countermsurgency strategies are defeated quickly or agree to peace terms early in the war.

⁵⁶ The sources consulted are too numerous to cite here. Among the more frequently cited reference sources consulted are Clodfelter, Micheal, Warfare and Armed Conflicts: A Statistical Reference to Casualty and Other Figures, 1500–2000 (London: McFarland, 2002)Google Scholar; The New York Times; The Encyclopedia Britannica Online; and the Library of Congress Country Study Series.

⁵⁷ Due to the unique geographical dispersion of the First and Second World Wars and due to the fact that different states or coalitions of states took on the dominant burden of the conflict in different regions, we separated these conflicts into several major battlefronts. Because this procedure tends to increase the significance of the world wars in our data set, as a robustness check we dropped all observations from the world wars and repeated our analysis. Our results remained consistent.

⁵⁸ In wars involving more than two states, we treat states as part of a coalition when the states coordinate their operations and aims at a broad strategic level but maintain significant autonomy mtactics at the operational level (for example, the allies in the First World War). We treat states as part of a “single state-led coalition” when there is an official unified command structure under the unilateral control of a single state, or when there is other tight control of all major operations by a single state (for example, the U.S.-led coalition in the 1991 Gulf War). In these cases, conflict-level variables such as troops committed and fatalities inflicted/incurred are coded for the coalition as a whole, but state-level attributes and characteristics such as regime type and treaty commitments are coded for the lead state only.

⁵⁹ Our results remained consistent when we dropped all cases of occupation from our analysis.

⁶⁰ Treaty ratification status was obtained from Roberts and Guelff (fn. 2).

⁶¹ For robustness, we also coded this variable based on unilateral ratification. Our results remained consistent.

⁶² Marshall, Monty G. and Jaggers, Keith, POLITY IV Data Set Users Manual (College Park: Center for International Development and Conflict Management, University of Maryland, 2000)Google Scholar. The polity categories of transition, interruption, and interregnum were filled with data from two primary sources: the democracy data set of Tatu Vanhanen, at http://www.sv.ntnu.no/iss/data/vanhanen; and Freedom House scores, at http://www.freedomhouse.org/ratings/index.htm.

⁶³ This variable was coded using multiple sources, especially Encyclopedia Britannica Online, http://www.britannica.com/. We also tested this hypothesis using separate variables representing racial and religious differences. Our results were unaffected.

⁶⁴ Coding rules for this variable were adapted from descriptions in Mearsheimer (fn. 49); and Stam (fn. 49).

⁶⁵ Unlike Stam (fn. 49), we do not code a separate strategy of “punishment” since we are interested in whether certain military strategies on the battlefield encourage states to adopt strategies of punishing the adversary.

⁶⁶ For example, in the Boer War of 1899–1902, the Boers began the war in October 1899 using a highly mobile conventional strategy against the British. In July 1900, however, the Boers switched to guerrilla warfare, a strategy they pursued until the war ended in May 1902. Since roughly 70 percent of the duration of the war was devoted to guerrilla warfare, the Boers received a coding of 70 for this variable.

⁶⁷ Ghosn, Faten, Palmer, Glenn, and Bremer, Stuart, “The MID3 Data Set, 1993–2001: Procedures, Coding Rules, and Description,” Conflict Management and Peace Science 21, no. 2 (2004)CrossRefGoogle Scholar.

⁶⁸ For observations in which data on military force size or military expenditures were missing for one or more parties in the war, the variable was coded utilizing the available measure. When both measures were missing (nineteen conflicts, twelve of which are occupations), we utilized the peak number of troops deployed by each side in the war. We did not adjust these measures for distance between states, since the figures represent the actual forces the state brought to bear in the conflict rather than the potential forces they might have deployed.

⁶⁹ To calculate the decay rate for military power over distance, we relied on the procedure described in Mesquita, Bruce Bueno de, The War Trap (New Haven: Yale University Press, 1981), 105–9Google Scholar.

⁷⁰ Population data were generated using EUGene. Missing data were filled in using multiple sources.

⁷¹ To test for robustness, we bootstrapped our regression coefficients. Bootstrapping is attractive as a check of robustness because it does not require the analyst to make any distributional assumptions. See Fox, John, Applied Regression Analysis, Linear Models, and Related Methods (Thousand Oaks: Sage, 1997), 494Google Scholar. We also estimated a censored regression model. All variables measuring strategic incentives remained positive and significant. This model, however, assumes that values less than zero on the dependent variable are possible but that all such observations were coded as zeros. This is not the case with civilian fatalities, since it is not possible to kill fewer than zero civilians.

⁷² Sheskin, David, The Handbook of Parametric and Nonparametnc Statistical Procedures (Boca Raton: Chapman and Hall, 2004), 267Google Scholar; King, Gary, “Variance Specification in Event Count Models: From Restrictive Assumptions to a Generalized Estimator,” American Journal of Political Science 33, no. 3 (1989), 764Google Scholar. For example, a negative binomial model better approximates the distribution of scores in sports such as baseball, soccer, and ice hockey (where each score is worth exactly 1 point) than for American football, in which touchdowns and field goals yield different point values. See Pollard, R., “Collegiate Football Scores and the Negative Binomial Distribution,” Journal of the American Statistical Association 68, no. 342 (1973)Google Scholar.

⁷³ Nevertheless, as an additional robustness check, we estimated a negative binomial model. Because negative binomial models require data that only assume integer values, it is not possible to correct for the skewed distribution of our dependent variable by logging. Therefore, we chose to drop several influential outliers from our analysis. Our results remained consistent with the OLS results reported below.

⁷⁴ We utilized the following baseline values to calculate the marginal effects: attrition/counterinsurgency strategy = 50 percent; war aims = 0; regime type -4.2 (median); treaty=l (mode); civilizational conflict = 0; duration = 223 days (median); relative capabilities .513 (median); population = 37,943,167 (median).

⁷⁵ Aronson, Elliot, Wilson, Timothy D., and Akert, Robin M., Social Psychology: The Heart and the Mind (New York: Harper Collins, 1994), 483.Google Scholar

⁷⁶ These findings diverge from those reported by Downes (fn. 29), although most of our other findings are consistent. It is difficult to identify the precise reasons for this discrepancy, however, since Downes employs a completely different data set in his analysis. The differences are too numerous to describe here, but the following examples represent three of the most significant differences. Most notably, Downes uses a dichotomous measure of “civilian victimization,” whereas we utilize a continuous measure of intentional civilian fatalities. Second, our measure of military strategy is also continuous—reflecting the percentage of the war devoted to different battlefield strategies—whereas Downes uses a dichotomous measure. Finally, Downes drops from his analysis combatants that he codes as lacking the ability to kill civilians, whereas we include all combatants and attempt to control for the balance of material capabilities between them. As described above, coding the capability to kill civilians is difficult to assess objectively and vulnerable to possible endogeneity. We believe that each of these choices give us greater confidence in our results.

⁷⁷ Coding for this variable was based on Huth and Allee (fn. 23).

⁷⁸ Results not recorded due to space constraints.

⁷⁹ Valentino (fn. 32), 17–22, 56–57.

⁸⁰ Dower (fn. 38).

⁸¹ Sherry, Michael, The Rise of American Air Power (New Haven: Yale University Press, 1987), 260Google Scholar, 314.

⁸² Gilbert, Martin, The Second World War (New York: Henry Holt, 1991), 680.Google Scholar

⁸³ This variable was logged to correct for a skewed distribution.

⁸⁴ Quoted in Grayling, A. C., Among the Dead Cities (New York: Walker and Company, 2006), 149.Google Scholar

⁸⁵ Churchill, Winston, The Second World War, vol. 2 (Boston: Houghton Mifflin, 1949), 643.Google Scholar

⁸⁶ Quoted in Glover, Jonathan, Humanity: A Moral History of the Twentieth Century (New Haven: Yale University Press, 1999), 82.Google Scholar

⁸⁷ Quoted in Koch, H. W., “The Strategic Air Offensive against Germany,” Historical Journal 34, no. 1 (1991), 134CrossRefGoogle Scholar.

⁸⁸ Alberto Gonzales, “Decision RE Application of the Geneva Convention,” Memorandum dated 2002; accessed online, National Security Archives, http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB127/02.01.25.pdf

⁸⁹ Martin (fn. 14); Simmons (fn. 14); Simmons, Beth A. and Hopkins, Daniel J., “The Constraining Power of International Treaties: Theory and Methods,” American Political Science Review 99, no. 4 (2005)CrossRefGoogle Scholar.

⁹⁰ Hathaway, , “Do Human Rights Treaties Make a Difference?” Yale Law Journal 111, no. 8 (2002)CrossRefGoogle Scholar.

⁹¹ Gonzales (fn. 88).

⁹² Carr, Caleb, The Lessons of Terror: A History of Warfare against Civilians (New York: Random House, 2002)Google Scholar; Pape (fn. 48); Horowitz, Michael and Reiter, Dan, “When Does Aerial Bombing Work? Quantitative Empirical Tests, 1917–1999,” Journal of Conflict Resolution 45, no. 2 (2001)Google Scholar.