Reining in Rebellion Reining in Rebellion Raúl L. Madrid and Luis L. Schenoni The Decline of Political Violence in South America, 1830–1929 N ineteenth-century South America was plagued by rebellion. The French political scientist Alexis de Tocqueville remarked in 1835 that “the turmoil of revolution is . . . the most natural state of the South American Spaniards at the present time.”1 Two decades thereafter, Bolivian President Manuel Isidoro Belzú summarized the plight of most statesmen in the region when he complained about “successive revolutions, revolutions in the south, revolutions in the north, revolutions fomented by my enemies, headed by my friends, put together in my house, arising from my side; holy God!”2 Although not all of these revolts escalated to full-scale civil wars, they collectively killed hundreds of thousands, generated continual political instability, devastated economies, and forestalled growth. By the turn of the century, however, South American countries began to experience signiªcantly fewer revolts. Indeed, the number of revolts dropped from an average of more than forty-ªve per decade in the nineteenth century to less than twenty-one per decade from 1900 to 1929. Revolts did not decline in all countries of the region: rebellions continued to occur frequently in Ecuador and Paraguay during the early twentieth century. Nor did all types of revolts diminish at the same rate. Whereas the number of revolts that originated outside the state apparatus were four times as frequent as military coups at the beginning of the nineteenth century, these two types of revolts occurred at roughly the same rate in the 1920s, owing to the sharp decline of the former. In addition, revolts in the ªrst few decades of the twentieth century Raúl L. Madrid is the Harold C. and Alice T. Nowlin Regents Professor in Liberal Arts at the University of Texas at Austin. Luis L. Schenoni is an Assistant Professor in the Department of Political Science at University College London. The authors would like to acknowledge the helpful comments from Jennifer Cyr, John Gerring, Wendy Hunter, Gerardo Munck, Kurt Weyland, and the anonymous reviewers on earlier versions of this article. Matt Martin and a host of undergraduate interns at the University of Texas at Austin provided crucial research assistance. Online supplementary materials for this article are available at https://doi.org/10.7910/DVN/WEMZ8V. 1. Alexis de Tocqueville, Democracy in America (1835; repr., New York: Alfred A. Knopf, 1945), 251. 2. Nicanor Aranzaes, Las revoluciones de Bolivia [Bolivia’s revolutions] (La Paz: Casa Editora Talleres Gráªcos, 1918), 158. According to Nicanor Aranzaes’s calculations, Bolivia had 185 revolts, mutinies, and coups from 1826 to 1903. International Security, Vol. 48, No. 3 (Winter 2023/24), 129–167, https://doi.org/10.1162/isec_a_00479 © 2024 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology. 129 International Security 48:3 130 tended to involve fewer people, to cause fewer casualties, and to end more quickly than they had in the nineteenth century. What led to this dramatic region-wide decline in revolts? Why did some types of rebellions diminish more quickly than others, and why did a couple of states fail to conform to the regional pattern? These questions are not just of historical interest. The decline of revolts had important long-term implications for development and democracy in the region. To begin with, the dramatic reduction in rebellions provided the political stability necessary for the sustained economic and social progress that took place in South America during the twentieth century. In addition, the decline in revolts helped strengthen constitutional rule and usher in an era of democracy in some South American countries. To be sure, most South American countries remained authoritarian during this period, and even those countries that democratized did not become fully democratic. Nevertheless, the decline in revolts laid the groundwork for the ªrst wave of democratization in the region. In Argentina, Chile, Colombia, and Uruguay, opposition parties abandoned the armed struggle and began to focus on the electoral path to power, pushing for democratic reforms that would level the electoral playing ªeld.3 Although military coups interrupted democratic rule in these countries in the decades that followed, the democratic experiences that they enjoyed during this period helped build enduring democratic institutions and practices. Existing scholarship pays surprisingly little attention to the decline of revolts in South America in part because of the absence of comprehensive data on the rebellions. Although historians have provided insightful analyses of the causes and consequences of revolts in individual countries,4 neither they nor social scientists have cataloged the decline of revolts in the region as a whole or identiªed and explained how this trend varied across countries and revolt 3. See Raúl L. Madrid, “The Partisan Path to Democracy: Argentina in Comparative Perspective,” Comparative Political Studies 52, no. 10 (2019): 1535–1569, https://doi.org/10.1177/001041401983 0738; Raúl L. Madrid, “The Origins of Democracy in South America” (unpublished manuscript, 2023). 4. See, for example, Aranzaes, Las revoluciones de Bolivia; Antonio Arráiz, Los días de la lra: Las guerras civiles en Venezuela, 1830–1903 [The days of wrath: Civil wars in Venezuela] (Valencia, Venezuela: Vadell Hermanos Editores, 1991); Carlos Camacho Arango, Margarita Garrido Otoya, and Daniel Gutiérrez Ardila, eds., Paz en la república: Colombia, siglo XIX [Peace in the republic: Colombia, nineteenth century] (Bogotá: Universidad Externado de Colombia, 2018); Pablo Camogli, Batallas entre hermanos: Todos los combates de las guerras civiles Argentinas [Battles between brothers: All the combats of the Argentine civil wars] (Buenos Aires: Aguilar, 2009); Rafael Pardo Rueda, La historia de las guerras [The history of the wars] (Bogotá: Ediciones B Colombia, 2004). Reining in Rebellion 131 type.5 Moreover, the main causes of the revolts that the literature has identiªed cannot easily explain their decline. Existing studies tend to emphasize the factors that motivated the rebels, such as ethnic and racial inequalities,6 religious cleavages,7 regionalist sentiments,8 electoral frustrations,9 and a lust for economic resources or power.10 These grievances and motivations persisted into the twentieth century, however, and thus they do not offer a compelling explanation for the long-term, gradual decline of rebellion across the region. This study adopts an alternative perspective, one that is more suited to capture the conditions that enable revolts. In line with the civil war literature, we maintain that military weakness encourages revolts from outside the state apparatus. Many of the South American revolts of the nineteenth century, we argue, stemmed from the weaknesses of the region’s militaries, but the expansion and professionalization of the armed forces at the end of the century discouraged subsequent revolts. These internationally driven military buildups indirectly enabled states in the region to achieve a monopoly on violence, which radically transformed domestic politics in South America. This study advances the civil war literature conceptually, empirically, and theoretically. Our central conceptual innovation is an original typology that identiªes four distinct categories of revolts on the basis of whether the rebel 5. The most comprehensive source on revolts in the region is Robert L. Scheina, Latin America’s Wars: The Age of the Caudillo, 1791–1899 (Washington, DC: Brassey’s, 2003). But Scheina neither explains the decline in revolts nor provides an overarching analysis. 6. Brooke Larson, Trials of Nation Making: Liberalism, Race, and Ethnicity in the Andes, 1810–1910 (New York: Cambridge University Press, 2004); Marta Irurozqui, “The Sound of the Pututos: Politicisation and Indigenous Rebellions in Bolivia, 1826–1921,” Journal of Latin American Studies 32, no. 1 (2000): 85–114, https://doi.org/10.1017/S0022216X99005477. 7. Luis Javier Ortiz Mesa, Ganarse el cielo defendiendo la religion: Guerras civiles en Colombia, 1840– 1902 [To gain heaven defending religion: Civil wars in Colombia, 1840–1902] (Bogotá: Universidad Nacional de Colombia, 2005); Rebecca Earle, “The War of the Supremes: Border Conºict, Religious Crusade or Simply Politics by Other Means,” in Rebecca Earle, ed., Rumours of Wars: Civil Conºict in Nineteenth-Century Latin America (London: Institute of Latin American Studies, University of London Press, 2000), 119–134. 8. David Bushnell, Reform and Reaction in the Platine Provinces, 1810–1852 (Gainesville: University of Florida Press, 1983); John Charles Chasteen, Heroes on Horseback: A Life and Times of the Last Gaucho Caudillos (Albuquerque: University of New Mexico Press, 1995). 9. Eduardo Posada-Carbó, “Elections and Civil Wars in Nineteenth-Century Colombia: The 1875 Presidential Campaign,” Journal of Latin American Studies 26, no. 3 (1994): 621–649; Carlos Malamud, “The Origins of Revolution in Nineteenth-Century Argentina,” in Earle, Rumours of Wars, 29–48. 10. Alejandro M. Rabinovich and Natalia Perea Sobrevilla, “Regular and Irregular Forces in Conºict: Nineteenth Century Insurgencies in South America,” Small Wars and Insurgencies 30, no. 4/5 (2019): 775–796, https://doi.org/10.1080/09592318.2019.1638538; Scheina, Latin America’s Wars. International Security 48:3 132 leaders originate inside or outside the state apparatus (insider or outsider revolts) and whether they are elites or of the masses. Our main empirical contribution is the development and analysis of a comprehensive database on revolts from 1830 to 1929.11 This database enables us to rigorously document the decline in overall revolts during this period and to identify important trends in different types of revolts that the literature overlooks. It also enables us to carry out what we believe is the ªrst quantitative analysis of South American revolts during this period. Our main theoretical contribution is to show that increases in the size and professionalization of the military reduce some types of revolts but not others. Ours is not the ªrst study to suggest that the strengthening of South American militaries at the outset of the twentieth century led to a decline in revolts. But it is the ªrst to systematically document this trend and to show how increased military strength reduced the number of revolts from outside the state apparatus, such as elite insurrections and popular uprisings, but not revolts from inside the state, such as military coups and mutinies. The decline in outsider revolts was highly beneªcial in part because, as we show, outsider revolts tended to be the largest and bloodiest rebellions. But the continuation of insider revolts undermined democracy and political stability in the region because they were the most likely to succeed in overthrowing the president.12 An additional theoretical contribution is to show that three exogenous factors led to the strengthening of the armed forces: the export boom, the threat of interstate conºict, and victories in foreign wars. For much of the nineteenth century, South American governments lacked the resources to invest extensively in their militaries. But when foreign trade expanded in the late nineteenth century, it generated new revenues that the states could use to import sophisticated weaponry, hire foreign military advisers, establish military schools, and expand the size of their armies. To exploit exportable resources, states also needed to effectively control distant corners of their territories, which heightened tensions with their neighbors and provided the motivation for upgrading the military. In addition, the intense militarized interstate disputes that persisted into the early twentieth century set off region-wide arms races as states rushed to respond to increased foreign threats. Finally, the major 11. Online supplementary materials for this article are available at https://doi.org/10.7910/ DVN/WEMZ8V. 12. This article focuses on explaining the onset of major revolts. We also discuss major revolts’ outcomes (e.g., how often the revolts overthrew a president) because they affected the likelihood of future revolts. Reining in Rebellion 133 interstate wars that South America experienced in the late nineteenth century also had a lasting impact on the region’s armed forces, strengthening or weakening them, depending on the war outcomes. To explore these questions, this study employs a multi-method strategy and harnesses an abundance of qualitative and quantitative evidence. We draw on numerous historical studies to build our database, to describe the main characteristics and ubiquity of revolts during the nineteenth century, and to show how these diminished once South American countries expanded and professionalized their armed forces. We then demonstrate that military strength can explain temporal and cross-national variation in outsider revolts better than any alternative hypothesis. Nineteenth-century South America provides an interesting and relatively unexplored laboratory to explore the causes of domestic conºict. The sheer length of the historical period, the lack of attrition of South American states (compared to European ones at the time), and the high intensity and frequency of revolts make for an ideal setting to test available theories.13 Explaining the decline of revolts in nineteenth-century South America may also shed light on changes (or lack thereof) in political violence in other time periods and regions. Military professionalization and state building came earlier in South America than it did in Africa and parts of Asia. That African and Asian countries developed their militaries relatively slowly may well explain why they remained prone to outsider revolts for much of the twentieth century. This article is organized as follows. The ªrst section presents a typology of revolts and shows how the frequency and type of revolts changed over time and across countries in South America. The second section discusses the existing literature on revolts in the region and on civil war more generally. It uses these literatures to develop an explanation for variation in the frequency of revolts that is focused on military strength. The third and fourth sections examine how military strength evolved in South America during the nineteenth and early twentieth centuries. They show how the strengthening and professionalization of the military reduced the frequency of certain types of revolts at the turn of the century. The ªfth section presents a statistical test of this argument, and the conclusion highlights the theoretical, conceptual, and empirical contributions of this study. 13. We exclude the decades before 1830 to omit conºicts associated with independence struggles and to capture a period of greater stability in the pool of states. Uruguay became a sovereign state in 1828, and Ecuador and Venezuela did so in 1830. International Security 48:3 134 The Decline in Revolts in South America Existing studies lack the data to precisely deªne the frequency of revolts in South America during the nineteenth and early twentieth centuries and to explain how they varied over time and across states. Indeed, the sheer number of revolts has led some scholars to despair of the possibility of counting them all.14 To develop a comprehensive count of revolts, we used more than 250 historical sources to construct an original database of all rebellions in the region from 1830 to 1929.15 We deªne a revolt as an instance when an identiªable domestic political group deªes the authority of the state by using or credibly threatening to use violence.16 Our datasets reveal a dramatic decline in major revolts in South America from the nineteenth to the twentieth century, as ªgure 1 indicates.17 We focus on major revolts, which we deªne as those that involved at least 500 rebels, because they have the most important consequences—that is, they are mostly likely to produce regime change or lead to signiªcant bloodshed and economic disruption.18 Moreover, data on them are more plentiful, which reduces measurement and identiªcation errors. From 1830 to 1899, there were on average 0.45 active major revolts per country/year, meaning that each country had almost a ªfty-ªfty chance of facing an important rebellion in any given year. By contrast, in the ªrst three decades of the twentieth century, this average declined to 0.21, or approximately a one-ªfth chance of experiencing a major rebellion in any country/year. While the decline is partly because of the longer duration of revolts in the nineteenth century, the ªnding also holds for revolt onsets: the average number of revolts per decade was thirty in the nineteenth 14. Miguel Angel Centeno, Blood and Debt: War and the Nation-State in Latin America (University Park: Pennsylvania State University Press, 2002), 61; Brian Loveman, For la Patria: Politics and the Armed Forces in Latin America (Lanham, MD: Rowman and Littleªeld, 1999), 43. 15. To identify and code revolts in South America between 1830 and 1929, we reviewed more than 250 historical works. The sources consulted included general histories and compendia of revolts as well as studies focusing on speciªc revolts. We coded numerous revolt characteristics including: the type of rebellion; the duration and location of the revolt; the identity, leadership, and aims of the rebels; the number of participants and casualties; and the outcome of each revolt. See the online appendix for a discussion of the methods used to construct this dataset and how it differs from existing conºict databases. 16. South American states only gradually consolidated their national boundaries during the nineteenth century. We deªne political conºicts as revolts if they defy the central authorities, even in the absence of consolidated territorial boundaries. 17. An analysis of minor revolts shows an even steeper decline and suggests that the trends for major rebellions apply to all rebellions in our dataset. 18. We use the terms revolts and rebellions interchangeably. Reining in Rebellion 135 Figure 1. Decline of Major Revolts in South America by Decade NOTE: Major revolts are those involving at least 500 rebels. All calculations are based on the Latin American Revolts Dataset available in the online supplementary materials. century compared to only ªfteen in the early twentieth century. Similar trends exist for especially lengthy, large, or impactful rebellions. As ªgure 1 shows, revolts that lasted for more than one year, that involved more than 5,000 rebels, and that led to the overthrow of the chief executive all declined dramatically during the early twentieth century, amounting to only a handful of cases by the 1920s. Figure 2 complements this picture, showing how the frequency of major revolts varied across countries and time. Argentina was the most rebellious country from 1830 to 1899, with over eight major revolts per decade. By contrast, Chile and Paraguay had the fewest major revolts during this period. The number of revolts declined during the ªrst three decades of the twentieth century in all South American countries except for Ecuador and Paraguay. In the other eight South American countries, and particularly in Argentina, Bolivia, International Security 48:3 136 Figure 2. Major Revolts in South America, 1830–1899 and 1900–1929 and Uruguay, the decline in revolts was dramatic. As we discuss below, these four cases show how exogenous shocks, such as wars and export booms, affect military strength and the likelihood of revolts. Not all types of revolts declined at the same rate. To explore variation across different types of revolts, we identify four categories of revolts on the basis of whether the leader of the revolt came from inside or outside the national state apparatus, and whether the rebel leader hailed from the elite or the masses.19 Henceforth, we refer to revolts with leaders from outside the state apparatus as outsider revolts or rebellions, and revolts from inside the state apparatus, including the military, as insider revolts or rebellions. Figure 3 depicts our typ19. Some revolts have multiple leaders with different backgrounds, but in classifying them, we focus on the characteristics of the paramount leader of each revolt. Thus, a revolt led by an elite member of the opposition is categorized as an elite insurrection, even if some military units and generals joined the revolt. Reining in Rebellion 137 Figure 3. A Typology of Revolts Based on the Origins of Their Leaders ology. Although it is based to a large degree on previously conceptualized revolt types, such as coups and civil wars, our typology offers a novel systematization that is more in line with historiographical work. For example, it identiªes an important type of revolt, elite insurrections, which were prominent in South America during this period but have not been conceptualized by political scientists.20 Alternative categorizations that focus on the consequences of the revolts cut across our typology. Civil wars, for example, typically refer to revolts by non-state actors that exceed a threshold of 1,000 battle deaths—any of our revolt types may become civil wars if they escalate, although outsider rebellions are more likely to do so.21 Following Jonathan Powell and Clayton Thyne, we deªne coups as “illegal and overt attempts by the military or other elites within the state apparatus to unseat the sitting executive.”22 The vast majority of coups originate in the mili20. Nicolás Somma ªnds that most nineteenth-century insurgencies in Chile, Colombia, Costa Rica, and Uruguay were led by powerful elites. See Nicolás M. Somma, “When the Powerful Rebel: Armed Insurgency in Nineteenth-Century Latin America” (PhD diss., University of Notre Dame, 2011), 1–8. Although elite insurrections largely disappeared in South America in the twentieth century, they continued in parts of Africa and Asia, so this category may also shed light on more recent rebellions. 21. James D. Fearon, “Why Do Some Civil Wars Last So Much Longer Than Others?,” Journal of Peace Research 41, no. 3 (2004): 275–301, https://doi.org/10.1177/0022343304043770. 22. Jonathan M. Powell and Clayton L. Thyne, “Global Instances of Coups from 1950 to 2010: A New Dataset,” Journal of Peace Research 48, no. 2 (2011): 252, https://doi.org/10.1177/0022343310 397436. International Security 48:3 138 tary, although coups may also be undertaken by high-ranking government ofªcials, such as cabinet ministers. A classic example of a nineteenth-century coup was when General José María Melo, an active-duty military ofªcer, overthrew the president of Colombia in 1854. We identify sixty-six major coup attempts from 1830 to 1929, providing valuable additions to existing coup datasets that start around 1946.23 Elite insurrections, which are revolts led by elites from outside the state apparatus, sometimes consist of local elites attempting to secede or protest policies, but even more frequently they involve opposition parties or politicians taking up arms to overthrow the government. A prominent example was the 1895 Liberal Revolution in Ecuador, in which Liberal Party forces under Eloy Alfaro, an opposition leader, overthrew the government. Elite insurrections were by far the most common type of major revolt from 1830 to 1929: we record 152 of them during this period. Popular uprisings refer to rebellions led by non-elites who are located outside the state. Examples include the 1927 uprising of indigenous people in Chayanta, Bolivia, and other indigenous revolts, violent labor protests, and slave rebellions. We identify 34 major popular uprisings during this period. Finally, there are revolts from within the state that are led by non-elites, such as mutinies of rank-and-ªle soldiers or noncommissioned ofªcers. These revolts were typically smaller. Because we record few instances of major mutinies, we drop this category from the descriptive statistics. We focus on the origins of leaders for conceptual and theoretical reasons. Conceptually, the origins of leaders provide a clear and appropriate criterion to categorize a revolt as an insider or outsider rebellion.24 Insider rebellions, such as coups and mutinies, have leaders from within the state apparatus, whereas outsider rebellions, such as elite insurrections and popular uprisings, have leaders that come from outside of the state apparatus. Theoretically, leadership origin matters for at least two reasons. First, knowing a leader’s origin 23. John J. Chin, David B. Carter, and Joseph G. Wright, “The Varieties of Coups d’État: Introducing the Colpus Dataset,” International Studies Quarterly 65, no. 4 (2021): 1040–1051, https:// doi.org/10.1093/isq/sqab058. Fabrice Lehoucq has developed a coup dataset for Latin America that goes back to 1900. See Fabrice Lehoucq, “The Coup Trap in Latin America” (unpublished manuscript, 2023). 24. In the civil war literature, all organizations confronting the military are usually assumed to be non-state actors. Consider, for example, a rather common case whereby an unsuccessful military coup leads to a protracted confrontation. The civil war literature typically classiªes such a case as a civil war once the conºict surpasses a battle-death threshold, and the conspirators are considered non-state actors confronting the military. This classiªcation could lead to counting a single rebellion twice: ªrst as a coup attempt, and then as a civil war. We suggest that this instance would be recorded more properly as an insider rebellion by a state actor. Reining in Rebellion 139 helps determine the likelihood that a revolt will succeed. Revolts led by insider elites are more likely to succeed because insider elites tend to have greater access to resources, including troops, weaponry, ªnancing, and the media. As ªgure 4 (top panel) indicates, from 1830 to 1929, almost 71 percent of coup attempts in South America overthrew the government, as opposed to only 30 percent of elite insurrections and 3 percent of popular uprisings.25 Second, the origins of the rebel leaders also affect the size and costs of the revolts. Whereas insider rebellions tend to be resolved quickly and with minimal bloodshed, outsider revolts are usually more prolonged and more violent. From 1830 to 1929, 21 percent of popular uprisings and 14 percent of elite insurrections in South America lasted more than one year, as opposed to 6 percent of coups (middle panel). Similarly, 29 percent of outsider revolts (i.e., popular uprisings and elite insurrections) but only 10 percent of coups led to more than 1,000 battleªeld deaths (bottom panel). Even more interesting for our purposes, disaggregating the types of revolts by the origins of their leaders helps to shed light on the decline in revolts from the nineteenth to the twentieth century. As ªgure 5 shows, this decline was driven by a sharp drop in the number of outsider revolts, especially elite insurrections. During most of the nineteenth century, there were three times more outsider rebellions than insider revolts, but by the 1920s their frequencies were roughly the same. In sum, our new database indicates that revolts in South America declined sharply during the early twentieth century, and the large, lengthy, and bloody internal conºicts that plagued the region during the nineteenth century mostly came to an end. We show that this decline varied somewhat across countries. Paraguay was the only country to experience an increase in major revolts during the ªrst few decades of the twentieth century—a case that we discuss further below. We also ªnd that insider and outsider rebellions differed greatly in terms of their average size, duration, level of violence, and success in overthrowing the government. Finally, we show that outsider revolts drove the decline in political violence, while insider revolts remained relatively stable. In the next section, we show how the divergent trends in insider and outsider revolts can be explained by the strengthening of South American militaries at the beginning of the twentieth century. We ªnd that military strength discouraged elite insurrections and popular uprisings but not military coups. 25. These ªgures should be viewed with some caution, however, since some unsuccessful coup attempts may not be reported in the historical literature. International Security 48:3 140 Figure 4. Characteristics of Major Revolts in South America by Type of Rebellion, 1830– 1929 NOTE: The data in the middle panel refer to the percentage of revolts that lasted more than 1 year (left), from 1 month to 1 year (middle), and less than 1 month (right). The data in the bottom panel refer to the percentage of revolts that involved more than 1,000 deaths (left), 100–999 deaths (middle), and fewer than 100 deaths (right). Reining in Rebellion 141 Figure 5. Frequency of Major Insider and Outsider Revolts in South America by Decade NOTE: Insider revolts refer to rebellions that originate within the state apparatus (e.g., coups and mutinies). Outsider revolts refer to rebellions by non-state armed groups (e.g., elite insurrections and popular uprisings). Explaining the Decline in Revolts The historical literature stresses that nineteenth-century revolts in South America were complex and had a wide variety of causes. Most of the historical literature focuses on the motivations of the rebels. Robert Scheina, for example, argues: The causes for wars in Latin America during the nineteenth century are numerous and create a vivid, plaid tapestry. . . . The most vivid threads have been the race war, the ideology of independence, the controversy of separation versus union, boundary disputes, territorial conquests, caudilloism, intraclass struggles, interventions caused by capitalism, and religious wars.26 26. Scheina, Latin America’s Wars, xxiii. International Security 48:3 142 Frank Safford identiªes ªve types of explanations for these revolts, including cultural factors, economic structures, ªscal weakness, changing power relations among elite groups, and conºicting ideologies and interests.27 Various scholars also focus on the electoral grievances of the rebels.28 Much of the general social science literature on political violence similarly focuses on the motivations of the rebels, or what the literature sometimes refers to as grievances and greed. The conºict literature, for example, extensively explores how economic factors,29 ethnic and religious cleavages,30 and regime types31 affect the likelihood of revolts. We believe that the motivations of the rebels are important, but they cannot fully explain long-term trends in South American revolts, which is the purpose of this article. On the theoretical side, these explanations do not indicate why citizens had the opportunity to rebel in the ªrst place—that is, they do not explain why insurgents were able to assemble their armies and fend off government troops, irrespective of their motivations. On the empirical side, these 27. Frank Safford, “The Problem of Political Order in Early Republican Spanish America,” Journal of Latin American Studies 24, no. S1 (1992): 83–97, https://doi.org/10.1017/S0022216X00023798; Frank Safford, “Reºections on the Internal Wars in Nineteenth-Century Latin America,” in Earle, Rumours of Wars, 6–28. 28. See, for example, Malamud, “The Origins of Revolution in Nineteenth-Century Argentina”; Eduardo Posada-Carbó, ed., Elections before Democracy: The History of Elections in Europe and Latin America (London: Palgrave Macmillan, 1996); Paula Alonso, Between Revolution and the Ballot Box: The Origins of the Radical Party in the 1890s (Cambridge: Cambridge University Press, 2000). During the nineteenth century, revolts were considered a legitimate response to electoral manipulation and other forms of despotism. Hilda Sabato, Republics of the New World: The Revolutionary Political Experiment in Nineteenth-Century Latin America (Princeton, NJ: Princeton University Press, 2018), 112–115; Rebecca Earle, “Introduction,” in Earle, Rumours of Wars, 3–4. 29. Christopher Blattman and Edward Miguel, “Civil War,” Journal of Economic Literature 48, no. 1 (2020): 45, https://doi.org/10.1257/jel.48.1.3; Curtis Bell, “Coup d’État and Democracy,” Comparative Political Studies 49, no. 9 (2016): 1170, https://doi.org/10.1177/0010414015621081. 30. Lars-Erik Cederman and Luc Girardin, “Beyond Fractionalization: Mapping Ethnicity onto Nationalist Insurgencies,” American Political Science Review 101, no. 1 (2007): 173–185, https:// doi.org/10.1017/S0003055407070086; Philip Roessler, “The Enemy Within: Personal Rule, Coups, and Civil War in Africa,” World Politics 63, no. 2 (2011): 300–346, https://doi.org/10.1017/S004388 7111000049; Nils-Christian Bormann, Lars-Erik Cederman, and Manuel Vogt, “Language, Religion, and Ethnic Civil War,” Journal of Conºict Resolution 61, no. 4 (2017): 744–771, https://doi.org/ 10.1177/0022002715600755; Joan Esteban, Laura Mayoral, and Debraj Ray, “Ethnicity and Conºict: An Empirical Study,” American Economic Review 102, no. 4 (June 2012): 1310–1342, https://doi.org/ 10.1257/aer.102.4.1310. 31. James D. Fearon and David D. Laitin, “Ethnicity, Insurgency, and Civil War,” American Political Science Review 97, no. 1 (2003): 84–85, https://doi.org/10.1017/S0003055403000534; Jonathan Powell, “Determinants of the Attempting and Outcome of Coups d’État,” Journal of Conºict Resolution 56, no. 6 (2012): 1035, https://doi.org/10.1177/0022002712445732; Håvard Hegre, “Toward a Democratic Civil Peace? Democracy, Political Change, and Civil War, 1816–1992,” American Political Science Review 95, no. 1 (2001): 33–48, https://doi.org/10.1017/S0003055401000119. Reining in Rebellion 143 approaches struggle to account for the dramatic decline in revolts that occurred at the outset of the twentieth century when authoritarian regimes, ethnic cleavages, electoral fraud, interstate rivalries, and economic hardships remained widespread. Another approach in the conºict literature focuses on the weakness of the state rather than on the motivations of rebels as the main cause of revolts.32 Stemming in part from the study of revolutions,33 this approach “has become the dominant explanatory paradigm in the civil war literature.”34 The weak state approach suggests that motivations for rebellion (grievances and greed) are widespread, but they tend to only result in signiªcant revolts where the state lacks the ability to prevent or suppress rebellions. Revolts occur, in the words of James Fearon and David Laitin, because “ªnancially, organizationally, and politically weak central governments render insurgency more feasible and attractive due to weak local policing or inept and corrupt counterinsurgency practices.”35 This literature, however, usually focuses on the subnational or cross-national level, and rarely analyses long-term regional patterns. Building on this approach, as well as on the work of historians, we focus on a speciªc dimension of state capacity: military strength. We deªne military strength not simply as the number of troops in the military, but also the degree of its professionalization—that is, the sophistication of the weaponry, training, and leadership that the military possesses. We argue that revolts occurred frequently during the nineteenth century because South American countries had small armies that were poorly equipped, trained, and led. Once these states expanded and professionalized their armed forces in the early twentieth century, the number of revolts in the region declined precipitously. Strong militaries could defeat uprisings before they became major revolts, but, even more importantly, military strength discouraged revolts. Would-be rebels were unlikely to revolt if they believed that the rebellions would be quickly suppressed by a powerful military. 32. Fearon and Laitin, “Ethnicity, Insurgency, and Civil War”; Cullen S. Hendrix, “Measuring State Capacity: Theoretical and Empirical Implications for the Study of Civil Conºict,” Journal of Peace Research 47, no. 3 (2010): 273–285, https://doi.org/10.1177/0022343310361838; James D. Fearon, Governance and Civil War Onset, World Development Report 2011 Background Paper (Washington, DC: World Bank, 2010), http://hdl.handle.net/10986/9123. 33. Theda Skocpol, States and Social Revolutions: A Comparative Analysis of France, Russia and China (Cambridge: Cambridge University Press, 1979). 34. Lars-Erik Cederman and Manuel Vogt, “Dynamics and Logics of Civil War,” Journal of Conºict Resolution 61, no. 9 (2017): 1997, https://doi.org/10.1177/0022002717721385. 35. Fearon and Laitin, “Ethnicity, Insurgency, and Civil War,” 75–76. International Security 48:3 144 To be sure, this is not the ªrst study to suggest that military expansion and professionalization reduced revolts in South America in the twentieth century.36 Nevertheless, we go well beyond existing studies in documenting how increased military strength led to the region-wide decline. In addition, we show that increased military strength explains not only why revolts diminished in South America from the nineteenth to the twentieth century, but also why this happened more rapidly in some countries than in others given that not all states expanded and professionalized their militaries at the same time or to the same degree. Equally important, growing military strength explains why some types of revolts decreased in South America at the outset of the twentieth century while others did not. The strengthening of the armed forces discouraged outsider revolts (e.g., elite insurrections and popular uprisings) because it provided the military with a greater number of troops and much more sophisticated weapons and training than the rebels had. Yet the advances in technological sophistication and size of the armed forces did not necessarily discourage revolts from inside the state apparatus, such as military coups. Indeed, efforts to strengthen the military sometimes empowered those who sought to carry out coups as well as those who opposed them. Although military professionalization is supposed to marginalize the military’s role in politics and establish clear civilian control over the military,37 it did not achieve these aims in South America. As Alfred Stepan argues, militaries in this region have traditionally been responsible for maintaining both internal and external security, which provided them with a rationale to intervene in politics.38 The armed forces overthrew civilian leaders not just to resolve perceived threats to national security but also to safeguard their own interests as well as those of allied political elites. According to Linda Alexander 36. Edwin Lieuwen, for example, argues that the strengthening and professionalization of the military “made it progressively more difªcult to launch rebellions without at least some support from the nation’s regular armed forces.” See Edwin Lieuwen, Arms and Politics in Latin America, rev. ed. (New York: Praeger, 1961), 29–31. Frank Safford, meanwhile, suggests that “as trade and government revenues expanded, Spanish American governments increasingly had the ªscal and therefore the military strength to sustain themselves.” See Safford, “The Problem of Political Order,” 97. Neither military professionalization nor the regional decline of revolts is the focus of either scholar’s analysis. 37. See Samuel P. Huntington, The Soldier and the State: The Theory and Politics of Civil-Military Relations (Cambridge, MA: Belknap Press of Harvard University Press, 1957). 38. Alfred Stepan focuses on Latin American militaries during the late twentieth century, but the military’s involvement in internal security dates to the nineteenth century. See Alfred Stepan, “The New Professionalism of Internal Warfare and Military Role Expansion,” in Alfred Stepan, ed., Authoritarian Brazil: Origins, Policies, and Future (New Haven, CT: Yale University Press, 1973). Reining in Rebellion 145 Rodríguez, “professionalization had the long-term effect of politicizing the armed forces to defend their corporate interest, which they identiªed as synonymous with those of the nation.”39 Military professionalization may have even encouraged some coups by enhancing the conªdence and autonomy of military ofªcers and persuading some ofªcers that they could do a better job of governing than civilian leaders.40 Increases in military budgets and personnel also increased the inºuence of the armed forces and the number of potential coup conspirators, thereby complicating coup-prooªng efforts. For all these reasons, insider revolts, in contrast to outsider revolts, did not decline signiªcantly in the wake of the professionalization of South American militaries. The strengthening of the armed forces in the nineteenth century was exogenously driven. As our historical narratives show, South American rulers strengthened their militaries in the late nineteenth century mostly to deal with foreign threats. South America experienced two major wars and various smaller ones during the late nineteenth century. Numerous unresolved border conºicts that began in this period persisted into the early twentieth century. Equally important, improved international economic conditions helped provide South American governments with the revenues necessary to expand their militaries and import weaponry and foreign military advisers. Although most South American governments were in dismal ªnancial shape and lacked the funds to invest in their militaries for much of the nineteenth century, the export boom of the late nineteenth and early twentieth centuries put these countries on a much more solid ªnancial footing. The export boom also created friction. European settlers expanded the agricultural frontier and capital moved into peripheral regions, leading South American states to seek to exert control over formerly remote border areas where exportable commodities were produced. In this way, economic conditions, as well as regional competition and conºict, triggered an arms race of sorts, putting pressure on South American countries to expand and professionalize their militaries. The result was a regional trend of military strengthening that led to the region-wide decline of outsider rebellions. 39. Linda Alexander Rodríguez, “Introduction,” in Linda Alexander Rodríguez, ed., Rank and Privilege: The Military and Society in Latin America (Wilmington, DE: Scholarly Resources, 1994), xiii. 40. Alain Rouquié, The Military and the State in Latin America (Berkeley: University of California Press, 1987), 102–104; John Samuel Fitch, The Armed Forces and Democracy in Latin America (Baltimore: Johns Hopkins University Press, 1998), 6–7. International Security 48:3 146 The Weak Militaries of Nineteenth-Century South America The weakness of South America’s militaries during the nineteenth century stemmed from a variety of factors, including: the small size of armies, rudimentary weaponry, the paucity of military discipline and training, and the politicization of the ofªcer corps.41 In addition, South American states decentralized security, creating militias that sometimes turned against the national military. All these shortcomings encouraged outsider revolts. South American governments could ill afford to invest in their militaries for most of the nineteenth century because they were starved for funds, especially foreign currency. The wars of independence disrupted trade and destroyed South American economies, and political instability combined with a lack of infrastructure and inefªcient policies slowed economic recovery in the decades that followed. Per capita gross domestic product (GDP) grew at a rate of less than 0.6 percent annually from 1820 to 1870 in South America.42 Meager economic growth severely constrained tax revenues, which in turn limited government spending. Although military expenditures were relatively low, they typically accounted for a large share of state spending, reducing the ability of South American governments to address other needs. After the wars of independence, South American governments reduced the size of their militaries to alleviate their ªscal burdens. Most armies remained quite small throughout the bulk of the nineteenth century, particularly compared to their European counterparts. Bolivia typically had fewer than 2,000 personnel in its army during the nineteenth century.43 The Colombian military never exceeded 4,000 soldiers before the 1880s, and it often had fewer than 2,000 personnel.44 According to Centeno, less than 0.5 percent of the population usually participated in the militaries of South American countries.45 When a foreign or domestic threat required it, militaries usually swelled, but 41. John J. Johnson, The Military and Society in Latin America (Stanford, CA: Stanford University Press, 1964); Loveman, For la Patria; Rouquié, The Military and the State; Sabato, Republics of the New World. This article focuses on South American armies rather than navies, because the former were the principal forces used to suppress revolts. 42. José Antonio Ocampo and Luis Bértola, The Economic Development of Latin America since Independence (Oxford: Oxford University Press, 2013), 62. 43. James Dunkerley, Orígenes del poder militar: Bolivia, 1879–1935 [Origins of military power: Bolivia, 1879–1935] (La Paz: Plural, 2003), 71. 44. Fernando López-Alves, State Formation and Democracy in Latin America, 1810–1900 (Durham, NC: Duke University Press, 2000), 138; James L. Payne, Patterns of Conºict in Colombia (New Haven, CT: Yale University Press, 1968), 120. 45. Centeno, Blood and Debt, 224–225. Reining in Rebellion 147 in a rather ad hoc manner. During wartime the military would sweep through urban neighborhoods and rural villages, press-ganging into service whatever able-bodied people they could ªnd. A popular saying of the time was: “If you want more volunteers, send more chains.”46 The troops’ wages were meager, the government sometimes fell into arrears on payments, and soldiers frequently deserted despite severe punishments for doing so.47 In addition, the troops received little training. As João Resende-Santos notes, “Prior to the 1880s, none of the regional militaries had a standardized system of enlistment, training, and reserves.”48 Soldiers came overwhelmingly from the poorest sectors of the population and typically had little education, if any. Most of the soldiers were illiterate and many were vagrants and even criminals. Colombia reported in 1882 that only 30 percent of its troops could read.49 Military ofªcers in South America also lacked proper training and organization during this period. According to Brian Loveman, the nineteenth-century armies “were not organized under an operational general staff, did virtually no planning for diverse military threats, carried out few military exercises, and were unprepared for sustained combat.”50 Army ofªcers rarely attended military schools. For example, the Argentine War Ministry reported that only thirty of its approximately 1,400 army ofªcers in 1893 had received advanced training or graduated from a military academy.51 Some South American governments founded military academies during the nineteenth century, but these academies typically operated irregularly, and their curricula were woefully outdated. Political connections, rather than military expertise, determined ascent in the ofªcer ranks.52 In many South American countries, widespread promotions led to an excess of ofªcers, particularly at the higher ranks. Bolivia, for example, had one general for every 102 soldiers and one ofªcer for every six soldiers in 1841.53 Venezuela’s ofªcer ranks were even more bloated: a census 46. Quoted in Johnson, The Military and Society in Latin America, 54. 47. Rouquié, The Military and the State, 65. 48. João Resende-Santos, Neorealism, States, and the Modern Mass Army (New York: Cambridge University Press, 2007), 121. 49. Malcolm Deas, “The Man on Foot: Conscription and the Nation-State in Nineteenth-Century Latin America,” in James Dunkerley, ed., Studies in the Formation of the Nation-State in Latin America (London: Institute of Latin American Studies, University of London Press, 2002), 92. 50. Loveman, For la Patria, 30. 51. Resende-Santos, Neorealism, States, and the Modern Mass Army, 122. 52. Johnson, The Military and Society, 52–53; Resende-Santos, Neorealism, States, and the Modern Mass Army, 122; Loveman, For la Patria, 42–43; George Philip, The Military in South American Politics (London: Croom Helm, 1985), chap. 4; Rouquié, The Military and the State, 64–65. 53. Scheina, Latin America’s Wars, 263; Dunkerley, Orígenes del poder militar, 18. International Security 48:3 148 of the state of Carabobo in 1873 counted 3,450 commissioned ofªcers, including 627 colonels and 449 generals, out of a population of 22,952.54 South American militaries also lacked sophisticated weaponry for most of the nineteenth century, relying on pointed weapons (e.g., the lance, the pike, the sword, and the machete) rather than ªrearms.55 Antonio Arraíz writes that during the revolts, “Combat took place in a series of personal encounters in which people attacked each other with lances, swords, bayonets, ªsts and whatever was at hand.”56 Both sides typically had some ªrearms, but these were primitive weapons with limited range and accuracy. Even when South American militaries did obtain more sophisticated weapons, they often had problems repairing and servicing them, and sometimes let them slip into rebels’ hands.57 During the nineteenth century, most South American governments reorganized and expanded civic guards or urban and provincial militias, which had existed since colonial times.58 These militias were less expensive to maintain than the regular army, but they did little to enhance the authority of the central state. First, militia members typically had little training or equipment. There were exceptions, however, such as in Brazil, where the state militias, especially those of São Paulo and Minas Gerais, gradually became better trained and armed than the federal army.59 The government usually required militiamen to provide their own weapons and training, but they did not own ªrearms and rarely performed drills. In the Río de la Plata region, for example, they only trained one or two days per month during peacetime.60 Second, militias could not be counted on to support the government. Indeed, they often formed the main base of rebel armies, which was particularly problematic given that in most countries the militia troops vastly outnumbered the army.61 In some cases, the militias were set up or expanded to coun- 54. Philip, The Military in South American Politics, 87. 55. Scheina, Latin America’s Wars, 427. 56. Arraíz, Los días de la ira, 151. For a vivid description of nineteenth-century warfare, see Rabinovich and Sobrevilla Perea, “Regular and Irregular Forces,” 786–791. 57. Arraíz, Los días de la ira, 157; Somma, “When the Powerful Rebel,” 236; Scheina, Latin America’s Wars, 427. 58. Sabato, Republics of the New World, 90–96. 59. Resende-Santos, Neorealism, States, and the Modern Mass Army, 124. 60. Rabinovich and Sobrevilla Perea, “Regular and Irregular Forces,” 784. 61. In Chile, for example, the regular army had only 3,000 troops, whereas the civic guard reached 60,000 troops in the 1850s before gradually declining. See James A. Wood, The Society of Equality: Popular Republicanism and Democracy in Santiago de Chile, 1818–1851 (Albuquerque: University of New Mexico, 2011), 86–88; Somma, “When the Powerful Rebel,” 398. Reining in Rebellion 149 terbalance the regular army. In Uruguay, for example, the Blanco Party built up a national guard to offset the Colorado Party–dominated army.62 Despite periodic efforts to centralize control, in most countries the militias remained under the leadership of provincial and local authorities and at times represented a direct threat to the national government.63 In Argentina, provincial militias typically supplied both the troops and the weapons that were used in revolts during the nineteenth century,64 and in Brazil the local militias of southern states singlehandedly sustained a ten-year campaign against the imperial army during the Ragamufªn War.65 In many rural areas, local caciques and caudillos controlled unofªcial militias, which often participated in rebellions and guerrilla warfare.66 It is not a coincidence that Chile and Paraguay, the two South American countries that had perhaps the highest coercive capacity during much of the nineteenth century, also had the fewest revolts. Chile experienced numerous revolts prior to 1860 and a civil war in 1891, but its military prowess, demonstrated in the War of the Paciªc (1879–1883) against Bolivia and Peru, discouraged most domestic rebels in the late nineteenth century. Chile developed a strong military during this period not by expanding its size, but by making early investments in tactics, weaponry, and foreign training. Ofªcers were sent to study in France beginning in the 1840s, and a small French training mission was contracted in 1858.67 Early on, Chile centralized control of its national 62. Selva López Chirico, El estado y fuerzas armadas en el Uruguay del siglo XX [The state and the armed forces in Uruguay in the twentieth century] (Montevideo: Ediciones de la Banda Oriental, 1985), 29–30; Somma, “When the Powerful Rebel,” 150. 63. Sabato, Republics of the New World, 98–99. The militias were not a threat to the central state’s authority everywhere in South America: militias rarely revolted in Chile, and in some countries, such as Paraguay, they were abolished early. 64. Ezequiel Gallo, “Argentina: Society and Politics, 1880–1916,” in Leslie Bethell, ed., The Cambridge History of Latin America, vol. 5 (Cambridge: Cambridge University Press, 1986), 379; Riccardo Forte, “Incertidumbre y determinación: Ransición liberal y construcción del poder coactivo del estado en México y Argentina (ca. 1855–1880) [Uncertainty and determination: Liberal transition and the construction of coercive power in Mexico and Argentina (ca. 1855–1880],” Anuario de historia regional y de las fronteras [Yearbook of regional and border history] 7, no. 1 (2002): 241, https://revistas.uis.edu.co/index.php/anuariohistoria/article/view/1473. 65. José Iran Ribeiro, “O fortalecimento do estado imperial a través do recrutamento militar no contexto da Guerra dos Farrapos [The strengthening of the imperial state through military recruitment in the context of the Ragamufªn War],” Revista Brasileira da história [Brazilian history review] 31, no. 62 (2011): 271, https://doi.org/10.1590/S0102-01882011000200014. 66. Rabinovich and Sobrevilla Perea, “Regular and Irregular Forces,” 785. 67. Hernán Ramírez Necochea, Fuerzas armadas y política en Chile, 1810–1970 [The armed forces and politics in Chile, 1810–1970] (Havana: Casa de las Américas, 1985), 39–40; Tommie Hillmon Jr., “A History of the Armed Forces of Chile from Independence to 1920” (PhD diss., Syracuse University, 1963), 76; J. Samuel Valenzuela, Democratización vía reforma: La expansión del sufragio en Chile International Security 48:3 150 guard, which played an important role in quashing rebellions as well as turning out votes for the ruling party.68 As discussed in the next section, the Chilean state achieved an even greater monopoly on violence when a much larger German military mission arrived in 1885. Paraguay also initially enjoyed relative political stability thanks to its considerable military strength. During the mid-nineteenth century, Paraguay developed one of the largest and strongest militaries in the region. The Paraguayan government imported massive quantities of weapons, overhauled troop training, and brought in foreign ofªcers, most notably the Hungarian Lieutenant Colonel Francisco Wisner von Morgenstern, to modernize and discipline its army.69 Paraguay even built up an important domestic arms industry. By 1864–1865, the Paraguayan Army had 30,000–38,000 troops, including thirty infantry regiments, twenty-three cavalry regiments, and four artillery regiments, as well as 150,000 reservists.70 The country’s military strength effectively discouraged revolts before the War of the Triple Alliance (1864–1870), also known as the Paraguayan War. In this war, however, the combined forces of Argentina, Brazil, and Uruguay destroyed the Paraguayan military. Consistent with our expectations, Paraguay was plagued by revolts in the decades that followed. Military Strengthening In the late nineteenth to early twentieth centuries, South American states undertook major efforts to strengthen their militaries, often with the assistance of foreign military missions. They expanded the size of their armies, upgraded their weaponry, established new military schools, adopted meritocratic criteria [Democratization through reform: The expansion of suffrage in Chile] (Buenos Aires: IDES, 1985), 182. 68. Sabato, Republics of the New World, 107, 10–11; J. Samuel Valenzuela, “Building Aspects of Democracy before Democracy: Electoral Practices in Nineteenth-Century Chile,” in Eduardo PosadaCarbó, ed., Elections before Democracy, 228–231. 69. John Hoyt Williams, The Rise and Fall of the Paraguayan Republic, 1800–1870 (Austin: Institute of Latin American Studies, University of Texas Press, 1979), 79, 110–111; Thomas L. Whigham, The Paraguayan War, vol. 1, Causes and Early Conºict (Lincoln: University of Nebraska Press, 2002), 182; Richard S. Sacks, “Historical Setting,” in Dennis M. Hanratty and Sandra W. Meditz, eds., Paraguay: A Country Study (Washington, DC: Library of Congress, 1988), 24. 70. Melinda Wheeler Cooke, “National Security,” in Hanratty and Meditz, Paraguay, 205; Juan Manuel Casal, “Uruguay and the Paraguayan War: The Military Dimension,” in Hendrik Kraay and Thomas L. Whigham, eds., I Die with My Country: Perspectives on the Paraguayan War, 1864– 1870 (Lincoln: University of Nebraska Press, 2004), 187. Reining in Rebellion 151 for recruiting and promoting ofªcers, and banned private arms imports and local militias. As a result, their military strength increased and outsider revolts declined signiªcantly, in both number and intensity, during the ªrst few decades of the twentieth century. The only countries that continued to have numerous outsider revolts were those with the weakest militaries: Ecuador and Paraguay. South American countries experienced a signiªcant amount of interstate conºict in the nineteenth century, which put pressure on their governments to build up their militaries. From 1820 to 1914, Latin American countries fought almost as many interstate wars as European countries did, and these wars lasted much longer and killed a signiªcantly greater percentage of the population than they did in Europe.71 The War of the Triple Alliance, with an estimated 290,000 casualties, was the bloodiest war of that period, with battle deaths exceeding even those of the Crimean War.72 The casualties for the other major South American war of this period, the War of the Paciªc (1879–1883), were similar to the average European conºict of the time. Although there were no major wars in the region from 1884 to 1929, numerous militarized conºicts persisted.73 David Mares reports that from 1884 to 1918 alone, South American countries had thirty-one militarized interstate disputes, in which military force was used, threatened, or displayed.74 K. J. Holsti notes that in the region “one sees patterns of peace and war, intervention, territorial predation, alliances, arms-racing, and power-balancing quite similar to those found in eighteenthcentury Europe.”75 These conºicts provided two types of exogenous shocks affecting military strength. First, the threat of war forced every country to expand, modernize, and often mobilize its armed forces. South American countries may not have risked annihilation in conºicts with their neighbors, but they certainly risked 71. Luis L. Schenoni, “Bringing War Back In: Victory and State Formation in Latin America,” American Journal of Political Science 65, no. 2 (2021): 408, https://doi.org/10.1111/ajps.12552. 72. Micheal Clodfelter, Warfare and Armed Conºicts: A Statistical Encyclopedia of Casualty and Other Figures, 1492–2015, 4th ed. (Jefferson, NC: McFarland, 2017), 180. Civil wars such as the American Civil War and the Taiping Rebellion claimed even more lives but were not international. 73. Steven Ross Ligon, “The Character of Border Conºict: Latin American Border Conºicts, 1830– 1995” (PhD diss., Catholic University of America, 2002); Paul R. Hensel, “One Thing Leads to Another: Recurrent Militarized Disputes in Latin America, 1816–1986,” Journal of Peace Research 31, no. 3 (1994): 281–297, https://doi.org/10.1177/0022343394031003004. 74. David R. Mares, Violent Peace: Militarized Interstate Bargaining in Latin America (New York: Columbia University Press, 2001), 77. 75. K. J. Holsti, The State, War, and the State of War (Cambridge: Cambridge University Press, 1996), 153. International Security 48:3 152 losing territory and lives.76 For this reason, once one country strengthened its military, its neighbors and rivals felt compelled to do the same. As ResendeSantos puts it, “Intensifying military competition and war, in turn, prompted a chain reaction of large-scale military emulation,” resulting in military modernization that was “of a scale, intensity and duration not previously known in the region.”77 Second, war outcomes had an independent effect on military strength because defeat in war typically resulted in military downsizing, which was often imposed by the winners. Victory in war frequently led to military expansion because the victorious armed forces typically gained a great deal of inºuence that enabled them to obtain increased resources. Of these two types of shocks, the threat of war had the most important and long-lasting effects because the threat of conºict was more pervasive in South America than actual war. Military strengthening was expensive, but the export boom of the late nineteenth through early twentieth centuries brought new revenues to South American governments.78 The real value of exports increased almost tenfold from less than $1.3 billion in the early 1870s to $12.4 billion in the late 1920s in constant (1980) U.S. dollars, thanks in part to infrastructure improvements, technological developments, more liberal economic policies, and growing world demand.79 At the same time, foreign investment ºowed into the region, increasing from $1.1 billion in 1880 to $11.2 billion in 1929.80 Foreign investment helped capitalize the export sector and build infrastructure, such as railroads and ports, which made the exports possible. Not only did the expansion of foreign trade and investment provide the foreign currency to pay for imported weapons and foreign military missions, but it also provided incentives to build up the military given that the export boom depended on the ability of South American states to control the areas where export commodities were produced. When these areas were controlled by rebels, most of the time it was 76. As Miguel Angel Centeno points out, no Latin American country disappeared after 1840 as a result of war. See Centeno, Blood and Debt, 8. 77. Resende-Santos, Neorealism, States, and the Modern Mass Army, 37. 78. State building, including the strengthening of the military, also contributed to the expansion of exports by delivering public goods such as infrastructure and political stability. 79. Ocampo and Bértola, The Economic Development of Latin America, 86, 97; John H. Coatsworth, “Economic and Institutional Trajectories in Nineteenth-Century Latin America,” in John H. Coatsworth and Alan M. Taylor, eds., Latin America and the World Economy since 1800 (Cambridge, MA: Harvard University Press, 1998), 39–42. 80. Ocampo and Bértola, The Economic Development of Latin America, 124. Reining in Rebellion 153 easier for the state to co-opt rural caudillos by allowing them free rein in their domains than to militarily subjugate them.81 When two states disputed sovereignty over the territory in question, however, the conºict typically led to military buildups and even war: the War of the Paciªc, for example, originated in a dispute between Bolivia and Chile over nitrate-rich lands in the Atacama Desert.82 Export booms not only generated the incentives to wrestle land from neighboring states but also fueled conºict by bringing miners, farmers, and speculators into far-ºung disputed areas. Military competition was more intense where the threats of war were more pressing and where resources were more readily available.83 Wealthier South American countries, especially those experiencing export booms, such as Argentina and Chile, could more easily afford to make large investments in their armed forces. Indeed, Argentina and Chile engaged in a formidable arms race in the late nineteenth to early twentieth centuries, with both countries importing increasingly sophisticated weaponry and nearly going to war on several occasions from 1898 to 1902.84 In the early twentieth century, Argentina and Chile even obtained dreadnoughts, the era’s most sophisticated type of warship. Territorially small and surrounded by foes, Chile was the ªrst country to modernize its military, hiring a German mission headed by Captain Emil Körner in 1885. Argentina, which had territorial disputes with Chile, responded by hiring military advisers in the 1880s, and in 1899 it, too, contracted with a German military mission. Bolivia and Peru, which continued to claim the land that Chile had conquered in the War of the Paciªc, responded in kind. Peru commissioned a French mission in 1895, bringing in thirty-three French ofªcers to teach in Peruvian military schools from 1896 to 1914.85 The Bolivian military also hired various foreign ofªcers to teach in its military schools dur81. Sebastián Mazzuca, Latecomer State Formation: Political Geography and Capacity Failure in Latin America (New Haven, CT: Yale University Press, 2021). 82. William Sater, Chile and the War of the Paciªc (Lincoln: University of Nebraska Press, 1985). 83. Resende-Santos, Neorealism, States, and the Modern Mass Army. 84. Ibid.; Luis L. Schenoni, Bringing War Back In: Victory, Defeat, and the State in Nineteenth-Century Latin America (New York: Cambridge University Press, 2024). The victories of Argentina and Chile in foreign wars during the late nineteenth century contributed to the arms race by strengthening their militaries, energizing nationalist sentiments, and stiffening their position regarding territorial disputes. 85. Frederick M. Nunn, Yesterday’s Soldiers: European Military Professionalism in South America, 1890–1940 (Lincoln: University of Nebraska Press, 1983), 114–117; Teodoro Hidalgo Morey et al., Historia general del ejército del Perú: El ejército en la República, Siglo XIX [General history of the army of Peru: The military in the republic, nineteenth century], vol. 5 (Lima: Comisión Permanente de Historia del Ejército del Perú, 2005), 349–352. International Security 48:3 154 ing the 1890s, and in 1905, its ªrst French military mission arrived, followed by a German mission in 1910. The foreign missions gradually spread from Chile and its neighbors to the other South American countries. Some of these countries, such as Brazil, Paraguay, and Uruguay, contracted with European missions or advisers. Others, like Colombia, Ecuador, and Venezuela, hired Chilean military advisers to teach the Prussian military model and sent their own military ofªcers to train in Chile.86 With the support of the foreign missions, most South American countries moved to expand the size of their militaries by enacting laws that mandated military service. Chile was again the pioneer, instituting universal obligatory military service in 1900.87 In response, Argentina enacted a similar conscription law in 1901, and by 1910 it could ªeld a standing force of 250,000 soldiers.88 Meanwhile, Uruguay doubled and Peru and Venezuela tripled the sizes of their respective armies.89 South American militaries also sought to improve the training of ofªcers and troops by opening new military institutes and adopting meritocratic criteria for promoting ofªcers. In Chile, Captain Körner revamped military training along Prussian lines: the government created highly selective military academies for junior ofªcers and noncommissioned ofªcers in 1887, and subsequently established specialized schools for the infantry, the cavalry, and engineering.90 In addition, 130 Chilean ofªcers were sent to Germany for further training from 1895 to 1913.91 The Argentine military similarly modeled its educational curriculum on Germany’s war academy, employing various German ofªcers as instructors and sending over 150 ofªcers to train in Germany.92 With the support of its Chilean mission, the Colombian govern- 86. Patricia Arancibia Clavel, ed., El ejército de los Chilenos, 1540–1920 [The Chilean Army, 1540– 1920] (Santiago: Editorial Biblioteca Americana, 2007). 87. Resende-Santos, Neorealism, States, and the Modern Mass Army, 135–138. 88. Ibid., 201–202; Nunn, Yesterday’s Soldiers, 128–129. 89. Richard Kinney Moore, “Soldiers, Politicians, and Reaction: The Etiology of Military Rule in Uruguay” (PhD diss., University of Arizona, 1978), 40; Peter F. Klarén, “The Origins of Modern Peru, 1880–1930,” in Bethell, ed., The Cambridge History, 601. 90. William F. Sater and Holger H. Herwig, The Grand Illusion: The Prussianization of the Chilean Army (Lincoln: University of Nebraska Press, 1999), 44. 91. Resende-Santos, Neorealism, States, and the Modern Mass Army, 138–141. 92. Ibid., 203–206; Robert A. Potash, The Army and Politics in Argentina, 1928–1945: Yrigoyen to Perón (Stanford, CA: Stanford University Press, 1969), 4; Warren Schiff, “The Inºuence of the German Armed Forces and War Industry on Argentina, 1880–1914,” Hispanic American Historical Review 52, no. 3 (1972): 436–455, https://doi.org/10.1215/00182168-52.3.436; Fernando García Molina, La prehistoria del poder militar en la Argentina: La profesionalización, el modelo Alemán y la Reining in Rebellion 155 ment established several institutions to train military ofªcers and adopted meritocratic criteria for promotion.93 Most South American countries also imported a massive amount of foreign weaponry during this period. In the 1890s, for example, Chile undertook a major purchase of Krupp artillery, along with 100,000 German Mauser riºes. It also signed contracts to import more German weapons worth 15 million German marks, planning to equip a standing army of 150,000 soldiers.94 In 1889, Argentina acquired 60,000 Mauser riºes and in 1894, when tensions with Chile were high, it purchased so much equipment that, according to one highranking military ofªcial, it could “burn half of Chile.”95 During the early 1900s, Brazil also purchased several hundred thousand Mauser riºes and Krupp cannons from the Germans.96 Uruguay imported Krupp cannons, Colt and Maxim machine guns, and enough Mauser and Remington riºes to arm 50,000 troops.97 Venezuela similarly strengthened its military by purchasing Mauser riºes, Krupp artillery, and Hotchkiss machine guns, among other weapons.98 South American governments also took steps to monopolize the use of force by restricting nongovernmental entities’ ability to import arms and by asserting control over or eliminating regional and private militias. These measures were also driven in part by foreign competition, which put pressure on military organizations to become more centralized and cohesive in order to predecadencia del régimen oligárquico [The prehistory of military power in Argentina: Professionalization, the German model and the decline of oligarchic rule] (Buenos Aires: Eudeba, 2010), 47–65. 93. Sater and Herwig, The Grand Illusion, 44. 94. Resende-Santos, Neorealism, States, and the Modern Mass Army, 134. 95. Gilberto Ramírez Jr., “The Reform of the Argentine Army” (PhD diss., University of Texas at Austin, 1987), 183; Resende-Santos, Neorealism, States, and the Modern Mass Army, 198. 96. Frank D. McCann, “The Formative Period of Twentieth-Century Brazilian Army Thought, 1900–1922,” Hispanic American Historical Review 64, no. 4 (1984): 746, https://doi.org/10.1215/ 00182168-64.4.737; Frederick M. Nunn, “Military Professionalism and Professional Militarism in Brazil, 1870–1970,” Journal of Latin American Studies 4, no. 1 (1972): 35, https://doi.org/10.1017/ S0022216X0000167X; Resende-Santos, Neorealism, States, and the Modern Mass Army, 252–253. 97. Somma, “When the Powerful Rebel,” 160; Milton I. Vanger, José Batlle y Ordoñez of Uruguay: The Creator of His Times, 1902–1907 (Cambridge, MA: Harvard University Press, 1963), 89, 95; López Chirico, El estado y fuerzas armadas en el Uruguay, 42. 98. Scheina, Latin America’s Wars, 248; Tomás Straka, “Guiados por Bolívar: López Contreras, Bolivarianismo y pretorianismo en Venezuela [Guided by Bolívar: López Contreras, Bolivarianism and praetorianism in Venezuela],” in Domingo Irwin and Frédérique Langue, eds., Militares y poder en Venezuela: Ensayos historicos vinculados con las relaciones civiles y militares Venezolanas [The military and power in Venezuela: Historical essays on Venezuelan civil-military relations] (Caracas: Universidad Católica Andrés Bello, 2005), 103; Eduardo C. Schaposnik, La democratización de las fuerzas armadas venezolanas [The democratization of the Venezuelan armed forces] (Caracas: Fundación Nacional Gonzalo Barrios, Instituto de Investigaciones Sociales, 1985), 20. International Security 48:3 156 vent autonomous forces from being co-opted by foreign foes and used as ªfth columns. Culminating the process of centralization of the armed forces that started during the War of the Triple Alliance, the Argentine government passed a law in 1880 that prohibited “provincial authorities from forming military forces.”99 It also dissolved the national guard and integrated it into the army as a reserve force, boosting its numbers by 65,000 troops.100 Countries that were further away from the intense competition of the Southern Cone were slower to centralize military power, but they eventually implemented similar reforms. The Colombian government initiated a program in the early 1900s to collect the many weapons that its citizens had stockpiled before and during the War of a Thousand Days (1899–1902). By 1909, this program had collected 65,505 guns and 1,138,649 bullets, making it more difªcult for potential rebels to arm themselves.101 Similarly, Venezuela restricted the extent of weapons available to private citizens and subnational states in the early twentieth century,102 and in 1919, it abolished state militias.103 Although the strengthening of South American militaries was mostly driven by foreign threats, it discouraged internal revolts because would-be rebels knew that they had little chance of prevailing over a properly manned and equipped professional military. In 1911, for example, some warlords belonging to the opposition Blanco Party in Uruguay sought to carry out a revolt. The leadership of their party blocked them, however, stating that the rebels would be at a “notorious disadvantage” given the strengthening of the military, which was evidenced by the disastrous failure of previous revolts.104 In 1917, 99. Hilda Sabato, “¿Quién controla el poder militar? Disputas en torno a la formación del estado en el siglo XIX” [Who controls military power? Disputes surrounding the formation of the state in the nineteenth century], in Oscar A. Moreno, ed., La construcción de la nación argentina, el rol de las fuerzas armadas: Debates históricos en el marco del bicentenario, 1810–2010 [Building the Argentine nation, the role of the armed forces: Historical debates within the framework of the bicentennial, 1810–2010] (Buenos Aires: Ministerio de Defensa, Presidencia de la Nación, 2010), 137. 100. Nunn, Yesterday’s Soldiers, 48. 101. Charles W. Bergquist, Coffee and Conºict in Colombia, 1886–1910 (Durham, NC: Duke University Press, 1978), 225; Ricardo Esquivel Triana, Neutralidad y ordén: Política exterior y militar en Colombia, 1886–1918 [Neutrality and order: Foreign and military policy in Colombia, 1886–1918] (Bogota: Pontiªcia Universidad Javeriana, 2010), 265; Adolfo León Atehortúa Cruz, Construcción del ejército nacional en Colombia, 1907–1930: Reforma militar y misiones extrajenras [The construction of the national army in Colombia, 1907–1930: Military reform and foreign missions], ed. César A. Hurtado (Medellin, Colombia: La Carreta Editores, 2009), 21. 102. Brian S. McBeth, Dictatorship and Politics: Intrigue, Betrayal, and Survival in Venezuela, 1908– 1935 (Notre Dame, IN: University of Notre Dame Press, 2008), 6, 79–80. 103. Howard I. Blutstein et al., Venezuela: A Country Study (Washington, DC: Government Printing Ofªce, 1985), 248; Schaposnik, La democratización de las fuerzas armadas, 21. 104. Milton I. Vanger, The Model Country: José Batlle y Ordoñez of Uruguay, 1907–1915 (Hanover, NH: University Press of New England, 1980), 151–152. Reining in Rebellion 157 the Blanco leader Basilio Muñoz persuaded the party to sign a pact with the government and compete in elections because armed revolt would be futile.105 In Colombia as well, the professionalization of the military at the outset of the twentieth century discouraged the revolts that had been commonplace during the nineteenth century. Many Liberals wanted to rebel in response to the widespread fraud in the 1922 elections, but General Benjamín Herrera, the Liberal leader and presidential candidate that year, dissuaded them in part because the country’s strengthened military gave them little hope of success.106 In some countries, opposition groups increasingly sought support within the armed forces for their revolts because they recognized it would be impossible to defeat a professional military on their own. In Chile, for example, the parliamentary opposition successfully pursued the backing of the navy when it revolted against the government of José M. Balmaceda in 1891. Similarly, the Radical Civic Union recruited supporters within the army in its revolts against the Argentine government in 1893 and 1905. As a result of the strengthening of the military, the likelihood of a South American country having an active outsider revolt in any given year fell from 0.37 during the 1830–1899 period to 0.14 during the ªrst three decades of the twentieth century. Revolts from outside the state apparatus declined in large part because non-state armed groups recognized that they had little chance of success against the professionalized militaries. Popular uprisings had always been highly unlikely to overthrow the government in South America and none did so after 1900, but elite insurrections also became increasingly unlikely to prevail. From 1900 to 1929, only ªve elite insurrections succeeded in overthrowing the government, compared to thirty-eight during the last seven decades of the nineteenth century. Moreover, four of the ªve successful elite insurrections that took place during the 1900–1929 period occurred in the South American countries with the weakest militaries, Bolivia, Ecuador, and Paraguay. By contrast, the strengthening and professionalization of the military did not lead to a concomitant reduction in insider revolts (e.g., military coups) because such revolts remained relatively likely to succeed. There were approximately six onsets of insider revolts per decade during the ªrst three decades of 105. Milton I. Vanger, Uruguay’s José Batlle y Ordoñez: The Determined Visionary, 1915–1917 (Boulder, CO: Lynne Rienner, 2010), 232. 106. Anthony P. Maingot, “Colombia: Civil-Military Relations in a Political-Culture of Conºict” (PhD diss., University of Florida, 1967), 165–166. International Security 48:3 158 the twentieth century, down only slightly from an average of seven per decade from 1830 to 1899. Many of these insider revolts succeeded in taking power, which encouraged military ofªcers to continue to undertake them. Indeed, military coups became more likely to succeed in the twentieth century, presumably because the professionalization of the military and the weakening of the private and regional militias strengthened the military’s hand. Military coups succeeded in overthrowing the president 81 percent of the time from 1900 to 1929, compared to 66 percent of the time from 1830 to 1899. In those nations with the strongest militaries, outsider revolts almost disappeared in the twentieth century, although insider revolts continued to occur occasionally. Partly because of its military buildup, Chile experienced no outsider revolts during the ªrst three decades of the twentieth century, although it did experience two military coups. Argentina had the most revolts of any South American country during the nineteenth century, but its enormous military buildup during and after the War of the Triple Alliance discouraged revolts in the twentieth century. The revolt of 1905 led by the Radical Civic Union party, which was quickly quashed, was Argentina’s only elite insurrection during the ªrst three decades of the twentieth century.107 Not all South American countries developed strong militaries during the early twentieth century. Ecuador and Paraguay, which were among the smallest and poorest of the South American countries, took only meager steps to professionalize their militaries during this period. Their armed forces remained politicized, fragmented, poorly trained, and underequipped.108 As a result, both countries continued to be plagued by revolts. Paraguay suffered the most revolts, experiencing seven elite insurrections and seven military coups from 1900 to 1929, several of which were successful. Overall, the number of revolt onsets and revolt-years more than tripled com107. Although the Radicals were an opposition party and their main leaders were not state ofªcials, they sought military support for their 1905 revolt as they had in previous rebellions. But the failure to obtain enough support from the armed forces doomed the rebellion. Argentina also experienced several major labor protests in the early twentieth century that turned violent, but these were easily repressed by the military as well. 108. Luis N. Bareiro Spaini, Las fuerzas armadas y su profesionalidad: Realidad y perspectivas (una interpretación nacional y regional) [The armed forces and their professionalism: Reality and perspectives (a national and regional interpretation)] (Asunción: Intercontinental Editora, 2008); Arancibia Clavel, La inºuencia del ejército Chileno, 267; John Samuel Fitch, The Military Coup d’État as a Political Process: Ecuador, 1948–1966 (Baltimore: Johns Hopkins University Press, 1977), 16; Remigio Romero y Cordero, El ejército en cien años de vida republicana, 1830–1930 [The army in one hundred years of republican life, 1830–1930] (Quito: Centro de Estudios Históricos del Ejército, 1991), 380–383; Paul H. Lewis, Political Parties and Generations in Paraguay’s Liberal Era, 1869–1940 (Chapel Hill: University of North Carolina Press, 1993), 133. Reining in Rebellion 159 pared to the nineteenth century. The explanation for this reversal is straightforward: the Paraguayan military was destroyed in the War of the Triple Alliance. Although Paraguay had some 40,000 soldiers before the war and mobilized 70,000 troops at the height of hostilities, by the time occupation forces left in 1876, its army had declined to a mere 400 soldiers.109 The conºagration also affected the country’s territory and demographics—some historians estimate that it lost half of its territory and up to 60–70 percent of its population,110 which severely hampered Paraguay’s capacity to exploit formerly lucrative yerba mate and timber industries.111 In the decades that followed, the country lacked the will and the resources to rebuild a severely factionalized military. According to Harris Gaylord Warren, during this period, “Paraguay’s armed forces were hardly sufªcient to maintain internal order.”112 As late as the 1920s, the country still lacked anything resembling a professional army because the government continued to politically manipulate promotions and assignments.113 Paraguay did not take important steps to strengthen its military until the mid-1920s, when a growing conºict with Bolivia that ultimately led to the Chaco War (1932–1935) prompted the Paraguayan government to purchase foreign weapons, reorganize its general staff, and hire ªrst a French and then an Argentine military mission.114 Although Ecuador downsized its military considerably after its defeat in the Ecuadorian-Colombian War (1863), a cacao boom helped the Ecuadorian government fund some efforts to professionalize its military in the early twentieth century.115 In 1899, Quito hired a Chilean military mission to train Ecuadorian ofªcers, created new military schools, and began to send ofªcers to Chile for training.116 The Ecuadorian military also made military service obligatory, enacted new laws governing promotions and salaries, and purchased military 109. Osvaldo Kallsen, Historia del Paraguay contemporáneo [History of contemporary Paraguay] (Asunción: Imprenta Modelo, 1983), 33. 110. Whigham, The Paraguayan War; Thomas L. Whigham and Barbara Potthast, “The Paraguayan Rosetta Stone: New Insights into the Demographics of the Paraguayan War, 1864–1870,” Latin American Research Review 34, no. 1 (1999): 174–186, https://doi.org/10.1017/S0023879100024341. 111. Schenoni, “Bringing War Back In,” 414. 112. Harris Gaylord Warren, Rebirth of the Paraguayan Republic: The First Colorado Era, 1878–1904 (Pittsburgh, PA: University of Pittsburgh Press, 1985), 31. 113. Lewis, Political Parties and Generations, 133. 114. Bareiro Spaini, Las fuerzas armadas y su profesionalidad, 76–77; Lewis, Political Parties and Generations, 142. 115. Only in the 1930s did Ecuador muster a force of 6,000 soldiers, equivalent to the size of the one that preceded the 1863 conºict. See Peter V. N. Henderson, Gabriel García Moreno and Conservative State Formation in the Andes (Austin: University of Texas Press, 2008), 85. 116. Arancibia Clavel, La inºuencia del ejército Chileno, 190–196. International Security 48:3 160 equipment from Chile, France, and Germany.117 Nevertheless, the reforms took a while to bear fruit, and Roberto Arancibia Clavel and Remigio Romero y Cordero suggest that the long-term inºuence of the Chilean mission was relatively superªcial.118 The Ecuadorian military remained highly politicized, and the government continued to promote, demote, and discharge senior Ecuadorian ofªcers on the basis of their personal and political afªliations.119 For ªve years beginning in 1908, the government signiªcantly reduced the military budget, slashing the size of the standing army. By 1913, military salaries were lower than those of civilian employees.120 The weakness of the military encouraged the opposition to continue to carry out rebellions, some of which were successful. Rebels overthrew the government in 1906 and in 1911, and nearly did so again in the bloody 1911–1912 civil war. The military also struggled to suppress a rebellion that ravaged the province of Esmeraldas from 1913 to 1916. It was only after 1916 that the Ecuadorian military established a monopoly on violence. In terms of their level of military professionalization, the other South American countries fell somewhere in between Argentina and Chile on the one hand and Ecuador and Paraguay on the other. Bolivia, Brazil, Colombia, Peru, Uruguay, and Venezuela all took signiªcant steps to expand and modernize their militaries, but they did so later and on a smaller scale than Argentina and Chile. Nevertheless, except for Bolivia, the militaries of these other South American states gradually became stronger and far exceeded those of Ecuador and Paraguay in terms of their capabilities. Consequently, these countries made at least some progress in reducing outsider revolts in the early twentieth century. The Determinants of Outsider Revolts: A Statistical Test In this section we provide a summary statistical test of the impact of military strength on outsider revolts using our original panel data on ten South American countries from 1830 to 1929. We are interested in exploring the fac117. Paco Moncayo Gallegos, Fuerzas armadas y sociedad [Armed forces and society], vol. 44, Biblioteca de ciencias sociales [The library of social sciences] (Quito: Corporación Editora Nacional, 1995), 155; Fitch, The Military Coup d’État, 15; Arancibia Clavel, La inºuencia del ejército Chileno, 212. 118. Arancibia Clavel, La inºuencia del ejército Chileno, 267; Romero y Cordero, El ejército en cien años, 380–383. 119. Fitch, The Military Coup d’État, 16. 120. Ibid.; Linda Alexander Rodríguez, The Search for Public Policy: Regional Politics and Government Finances in Ecuador, 1830–1940 (Berkeley: University of California Press, 1985), 225. Reining in Rebellion 161 tors that affect the number of outsider revolts in a given country-year during this period.121 Because our outcome of interest is a count variable, we follow established procedure and use a series of Poisson regressions with two-way ªxed effects and clustered standard errors.122 The online appendix also includes some robustness checks that we performed. Because we are particularly interested in the impact that military strength and professionalization had on outsider revolts, we measure these factors in three different ways. First, we use data on a variable (milper) for the number of military personnel (in thousands) from the index of national material capabilities of the Correlates of War project.123 Second, we include data on a measure compiled by Nathan Toronto for the number of military academies in each country.124 Third, we employ data on a variable (v2stcritapparm) from the Varieties of Democracy project on appointment decisions in the armed forces.125 This variable ranges from 0 (none of the appointments are based on skill and merit) to 4 (all of them are). The coverage of these variables is slightly more limited from 1830 to 1845 for most countries, resulting in an unbalanced panel. Except for Uruguay, all countries enter the panel by 1854, and no observations drop because of attrition after a country enters the sample. When we include confounders, missingness follows the same pattern but is very marginal, forcing us to drop only ªfteen early observations. Model 1 includes these military variables, uses two-way ªxed effects to control for time and country invariant confounders, and reports standard errors clustered by country. According to our theory, however, these military variables are related to other variables (e.g., economic growth and interstate conºict) that can also shape the likelihood of revolts. It is, therefore, key to control for them and model 2 does precisely this, including potential time-variant and country-variant confounders. Since export booms can affect the size and quality of the military, as well as the propensity of outsiders to rebel, we in121. Country-year is a unit of analysis commonly used in statistical analyses of cross-national data in which each year that occurs in a country under study represents a separate observation. Because we are examining 10 countries during a 100-year period (1830–1929), our analysis contains 1,000 country-years. 122. Joshua D. Angrist and Jörn-Steffen Pischke, Mostly Harmless Econometrics: An Empiricist’s Companion (Princeton, NJ: Princeton University Press, 2009). 123. National Material Capabilities v6.0 dataset, Correlates of War, 2020, https://correlatesofwar .org/data-sets/national-material-capabilities. 124. Nathan W. Toronto, “Why Professionalize? Economic Modernization and Military Professionalism,” Foreign Policy Analysis 13, no. 4 (2017): 854–875, https://doi.org/10.1111/fpa.12093. 125. V-Dem Data Set, version 10, Varieties of Democracy, 2020, https://www.v-dem.net/data/ dataset-archive/. International Security 48:3 162 clude data on a variable from Giovanni Federico and Antonio Tena-Junguito that measures total exports in current U.S. dollars.126 Relatedly, the expansion of railroads and telegraphs might have facilitated both economic growth and military recruitment, and they might have increased the reach of state authorities, thereby narrowing opportunities to rebel. We therefore account for the hundreds of miles of railway track and telegraph lines in each country.127 To measure the potential impact of interstate conºict, we include a yearly count of each state’s involvement in militarized interstate disputes,128 as well as a dummy variable capturing whether the country lost a foreign war in the past ªfteen years.129 In addition, we include a series of controls that are common in the political violence literature. To control for the effect of hybrid regimes on political violence we use the Electoral Democracy Index (v2x_polyarchy) from the Varieties of Democracy project and its squared term.130 We also use data on an urbanization rate variable (e_miurbani) and the log of the population from the Varieties of Democracy project, given that outsider revolts and many of the aforementioned variables (e.g., military size) would presumably be affected by socioeconomic modernization and population size.131 Finally, we include the years elapsed since independence—and drop year ªxed effects—in model 3 to test if revolts declined simply as a function of time. Model 4 shows the robustness of our results to an important remaining confounder: GDP per capita. GDP per capita is perhaps the most signiªcant predictor of political violence in the literature, and we therefore include a measure 126. Giovanni Federico and Antonio Tena-Junguito, “World Trade, 1800–1938: A New Data-Set,” Working Paper 93, European Historical Economics Society, 2016, https://EconPapers.repec.org/ RePEc:hes:wpaper:0093. 127. Arthur S. Banks and Kenneth Wilson, Cross-National Time-Series Data, Databanks International, 2014, https://www.cntsdata.com. 128. Glenn Palmer et al., “The MID5 Dataset, 2011–2014: Procedures, Coding Rules, and Description,” Conºict Management and Peace Science 39, no. 4 (2022): 470–482, https://doi.org/10.1177/ 0738894221995743. 129. Schenoni, “Bringing War Back In.” 130. V-Dem Data Set. 131. The scarcity of data for nineteenth-century South America precludes controlling for other potential confounders. For example, there are no comprehensive time-series data on inequality or economic performance for this period. Nor are there reliable time-series data on the time-varying ethno-racial composition of South American countries or on the relative strength of Liberal and Conservative Parties during the nineteenth and early twentieth centuries. Nevertheless, our use of country ªxed effects should control for most of these unobservable characteristics that change slowly over time. The roughness of the terrain, which is also a prominent confounder in the civil war literature, is one of these time invariant factors. Similarly, our year ªxed effects should control for international shocks that affected all countries equally (e.g., commodity prices and global ªnancial crises). Reining in Rebellion 163 of it in real 2011 dollars (cgdppc) from the Maddison Project Database.132 Data are missing for numerous country-years, however, so model 4 should be viewed with some caution.133 Table 1 presents the results. In almost all models, the size of the military, the number of military academies, and the extent to which appointment decisions in the armed forces are meritocratic each has a negative and statistically signiªcant relationship with the number of outsider revolts. The only exception is model 4. When GDP per capita is included in the analysis, the number of military academies ceases to be signiªcant, but this could be explained by the reduced number of observations in this model. Most of the other variables have the negative signs that we expect, but they are not statistically signiªcant at the conventional 95 percent level.134 Figure 6 presents the same results in odd ratios. The chances of experiencing a new major outsider revolt in South America decreased by 30 percent with every additional 10,000-soldier increase in the size of the national military, as well as with every substantive increase in the meritocracy of the military (i.e., one point on the scale used), and with every new military academy. Overall, this statistical analysis of the determinants of revolts in South America from 1830 to 1929 provides support for the argument that military size and professionalization reduced the prevalence of outsider revolts but not of insider revolts such as coups and mutinies. As expected, when we switch the dependent variable from major outsider revolts to major insider revolts—i.e., coups—the military variables lose statistical signiªcance in most of the models.135 Although military strength discouraged regime outsiders from mounting rebellions, it clearly did not have the same impact on regime insiders. 132. Jutta Bolt et al., “Rebasing ‘Maddison’: New Income Comparisons and the Shape of LongRun Economic Development,” GGDC Research Memorandum 174 (Groningen, the Netherlands: University of Groningen, 2018), https://www.rug.nl/ggdc/html_publications/memorandum/ gd174.pdf. 133. We interpolated missing years for the gross domestic product (GDP) per capita variable, but in some cases the absence of data prevented interpolation, resulting in the nonrandom loss of 80 observations. In model 4, the number of observations declines to 695 and Paraguay drops out of the analysis altogether because it lacks GDP data for the entire 1830–1929 period. 134. The minimal change in the R-squared statistics when confounders are included in model 2 suggests that military variables explain most of the variance in the outcome. With observational data, endogeneity will inevitably remain a concern, but this ªnding should be taken as a strong indication that military strength might be mediating the impact of more structural geopolitical and economic variables, just as our historical review of the South American case suggests. Further model speciªcations conªrm this intuition. For example, when we drop all three military variables, one of the confounders, total exports, becomes signiªcant (at p ⬍ 0.05). This outcome suggests that the effect of export booms is mediated by military strength. For the results of these models and further information, see our online supplementary materials. 135. Results that are not shown are available upon request. International Security 48:3 164 Table 1. Determinants of Outsider Revolts in South America, 1830–1929 (1) (2) (3) Military personnel (in 10,000 soldiers) ⫺0.358* (0.16) ⫺0.368* (0.17) ⫺0.350* (0.16) ⫺0.309* (0.123) (4) Number of military academies ⫺0.390* (0.16) ⫺0.427* (0.21) ⫺0.381* (0.18) ⫺0.299** (0.09) Military appointments by skills and merit ⫺0.549* (0.23) ⫺0.581** (0.23) ⫺0.336* (0.16) ⫺0.395 (0.32) Urbanization rate ⫺0.049 (0.56) 0.032 (0.47) ⫺0.522 (0.55) V-Dem Electoral Democracy Index ⫺0.466 (0.38) ⫺0.320 (0.44) ⫺0.534 (0.38) V-Dem Electoral Democracy Index (squared) 0.106 (0.09) 0.092 (0.10) 0.121 (0.10) Militarized interstate disputes 0.178 (0.17) 0.178 (0.16) 0.194 (0.16) Defeat in foreign war (in previous 15-year period) ⫺0.106 (0.26) 0.166 (0.42) ⫺0.137 (0.29) Total exports ⫺0.178 (0.33) ⫺0.666 (0.54) ⫺0.150 (0.31) 0.279 (0.21) 0.095 (0.23) 0.353 (0.24) Hundreds of miles of railway track ⫺0.052 (0.12) 0.067 (0.15) ⫺0.091 (0.13) Population (log) ⫺0.005 (0.09) 0.008 (0.08) 0.060 (0.11) Hundreds of miles of telegraph lines Years since Independence ⫺0.010 (0.01) GDP per capita ⫺0.000 (0.00) constant ⫺3.761 (6.34) Pseudo R-squared 0.2252 ⫺1.317 (13.99) 0.2344 12.938 (6.96) 0.1371 ⫺7.160 (8.62) 0.2540 Fixed effects two-way two-way country two-way Standard errors clustered clustered clustered clustered 800 775 775 695 N NOTE: Poisson regressions on number of revolts per country-year. Standard errors in paren- theses. Country- and year- dummy variables not shown. Model 1 includes only military variables; models 2–4 include a range of control variables. * p ⬍ .05; **p ⬍ .01; ***p ⬍ .005 Reining in Rebellion 165 Figure 6. Determinants of Outsider Revolts in South America NOTE: Coefªcient plot with conªdence intervals based on the results from the four models in table 1 (represented here in odds ratios). Conclusion This paper provides the ªrst systematic cross-national analysis of the causes of the dramatic decline in revolts that occurred in South America from the nineteenth to the twentieth century. We show that the expansion and professionalization of the military signiªcantly reduced revolts by political outsiders in the region. The importance of this decline is clear: it vastly reduced the number of lives lost to violence, brought greater political stability to the region, and International Security 48:3 166 helped pave the way for a lengthy period of economic growth and state building. Military professionalization also laid the groundwork for the ªrst wave of democratization in the region by encouraging opposition parties to abandon the armed struggle and focus on the electoral path to power. Increased military strength did not, however, reduce insider revolts such as military coups, which continued to undermine political stability and eventually democracy in the region. Surprisingly, there has been relatively little cross-national research into the causes of this major historical turn or, for that matter, other region-wide declines in internal political violence. By providing comprehensive data on these rebellions and a systematic analysis of their causes, this article seeks not only to shed light on a critical juncture in South American political development, but also to stimulate more research on such regional dynamics. The ªndings of our study are consequential for the political science literature on conºict, which remains largely segmented into analyses of coups, civil wars, and similarly rigid and narrow categories. Our long-term historical analysis, which draws on the work of historians, suggests that political scientists might want to consider revolts as a broader category of political violence, one that disregards battle-death thresholds and political goals and comprises all instances of the use or threat of violence by political groups that defy the authority of the state. We have demonstrated the feasibility and beneªts of this approach. Our inclusive typology and comprehensive coverage of revolts enable us not only to grasp the full extent of political violence in South America, but also illuminate a range of theoretically insightful points about the distinct causes of speciªc revolt types. Our historical study of South America also highlights a new category of revolts—rebellions led by elites from outside the state—that were widespread in the region during the nineteenth century. This article brings to the conºict literature one hundred years of history in a region equivalent to a considerable portion of the international system at the time. This amounts to a major empirical contribution to the study of political violence in general. Future research could use our newly generated dataset to further explore the causes and consequences of revolts. Scholars could employ the database to examine the determinants of insider rebellions, such as coups and mutinies. Similarly, they could analyze under what circumstances revolts lead to the overthrow of presidents or to large numbers of casualties. By highlighting the importance of military strength in the decline of revolts, this article also provides a valuable addition to the growing body of literature that attributes political violence to the deªciencies of the state. It does so by of- Reining in Rebellion 167 fering a compelling historical narrative and by demonstrating statistically that our measures of military size and professionalization outperform other typical explanations of political violence, including the usual measures of state infrastructural capacity. The causal pathway that our theory lays out (in which exogenous international shocks, such as export booms and interstate conºict, lead to the diffusion of military strength throughout the region) also provides a compelling explanation for a relatively understudied phenomenon: the simultaneous decrease of political violence throughout a region. We hope our approach will inspire others to examine whether similar patterns of conºict are present in other regions and time periods as well. Indeed, the small size and lack of professionalization of the armed forces can presumably help explain why newly independent countries in Africa and Asia were plagued by outsider revolts after independence in the twentieth century. It can also explain why these rebellions decreased abruptly when geopolitical pressures compelled states to strengthen their militaries.