Militarization, Democracy, and Development

Do Third World countries benefit from having large militaries, or does this impede their development? Kirk Bowman uses statistical analysis to demonstrate that militarization has had a particularly malignant impact in this region. For his quantitative comparison he draws on longitudinal data for a sample of 76 developing countries and for 18 Latin American nations.

To illuminate the causal mechanisms at work, Bowman offers a detailed comparison of Costa Rica and Honduras between 1948 and 1998. The case studies not only serve to bolster his general argument about the harmful effects of militarization but also provide many new insights into the processes of democratic consolidation and economic transformation in these two Central American countries.

“This book comes at a particularly appropriate moment, one in which the United States is rethinking its unconditional support for democratic regimes and may be moving toward support for almost any regime that will join it in its war against the terrorists. Bowman shows that this may prove to be a Faustian bargain, one with serious long-term consequences for development in the Third World. The quantitative and qualitative evidence in this work is very persuasive and should be troubling for those who support the view that ‘the enemy of my enemy is my friend.'”—Mitchell A. Seligson, University of Pittsburgh

“Bowman challenges the time-worn thesis that military buildup is good for growth by showing that militarization has had negative consequences on democracy, growth, and equity in Latin America. Is Bowman insinuating that if a country wants to grow, build democracy, and restore equity, it should abolish the armed forces? This is indeed a tantalizingly provocative thesis. . . . Bowman has written the most lucid yet controversial and polemic book on the military—must reading for anyone interested in this topic.”—E. Pang, Choice

“Combining many different types of analysis, Militarization, Democracy, and Development is a considerable accomplishment.
Bowman grapples with theoretically important issues, presents skillfully conducted and original empirical work, and provides a clear explanation for disparate patterns of political development in Cold War Central America. This book will not end the debates about the causal relationships it explores, but it is a solid contribution to them.”—Anthony W. Pereira, Perspectives on Politics

“Recent political science literature has produced few works that are as compelling and important as Kirk Bowman’s Militarization, Democracy, and Development. Bowman’s pathbreaking work makes significant contributions to important debates in the areas of social science methodology, the role of the military in Latin America, and the nature of Costa Rican and Honduran political development. Its magnificent research design and jargon-free presentation should make this book required reading in any advanced course dealing with Latin American politics or research methodology. . . . This beautifully organized and crafted work is social science at its best.”—Donald Share, The Americas

“This book is the best analysis of the relationship between militarization and development in modern Latin America ever published. . . . His work offers a model of how to combine quantitative and qualitative research, illustrating their different but complementary strengths. It also significantly advances our understanding of the effects of militarization, the political history of Costa Rica and Honduras, and the importance of historical events in promoting or hindering development. All told this excellent book deserves to be read by a wide spectrum of scholars in the field of macrosociology.”—James Mahoney, American Journal of Sociology

“Although at times it requires stopping to think about how everything relates back to the hypothesis, it is also refreshing to read an author who so obviously has mastered the literature, dived headfirst into archives, and cares deeply about the argument itself.”—Gregory Weeks, Latin American Studies

Kirk S. Bowman is Assistant Professor at The Sam Nunn School of International Affairs, Georgia Institute of Technology.

Contents

Part I: Introduction

1. Militarization and Development: The Research Question and the Research Design

2. Militarization: The Causal Variable, the Literature, and the Theory

Part II: Bullets vs. Ballots: Militarization and Democracy

3. Taming the Tiger: A Quantitative Analysis of Militarization and Democracy in Latin America

4. When Ballots Trump Bullets: Demilitarization and Democratic Consolidation in Costa Rica

5. When Bullets Trump Ballots: Militarization and Democratic Collapse in Honduras

Part III: Guns vs. Butter: Militarization, Economic Growth, and Equity

6. Guns vs. Butter: A Quantitative Analysis of Militarization and Material Development

7. Escaping the Lost Decade: Militarization and Economic Growth in Costa Rica and Honduras

PART IV: Summation

8. Conclusion

1 Militarization and Development

This book examines the relationship between militarization and development in Latin America. I assess the effect of military size and budgets on three separate indicators of development—democracy, economic growth, and equity. I review past research and identify and solve a policy–relevant and theoretically important intellectual puzzle: while the conventional wisdom and case study evidence conclude that large militaries are negatively associated with development, a large body of cross–national research finds that lots of soldiers are good for economic growth and equity in developing countries. Why?

I use a simple typology of militaries to generate an explanation for the different effect of militarization on development in Latin America than in other regions of the developing world. I posit that the conservative historical trajectory of Latin American militaries, the Cold War influence of the United States, and, above all, the considerable internal focus of the region's armed forces combined to produce a particularly malignant pattern of militarization in the region. I expand the explanation using concepts of state capacity and priorities or focus. I test three hypotheses using several types of data and both quantitative and comparative historical research methods. I start at the general level, using statistical techniques and longitudinal data to test the effect of militarization on democracy, economic growth, and equity in both a universal sample of developing countries and in Latin America. This exercise demonstrates that the different military missions, history, and external threat–levels result in a very different militarizationdevelopment relationship in Latin America than in a universal sample. The results are stark and unambiguous: militarization has a substantial and significant negative effect on democracy, economic growth, and equity in Latin America.

Finally, I present a comparative historical analysis of the effect of militarization on development in Costa Rica and Honduras in the 1948–98 period. The comparative historical chapters establish agency and sequence in the relationship and illuminate the causal mechanisms. They also clearly reveal the serious opportunity costs of militarization and show how large and powerful militaries can undermine state capacity and trump efforts for greater economic growth, equity, and democracy. These chapters also reveal many novel insights about the democratic consolidation process and economic transformations in Costa Rica and Honduras in the 1948–98 period.

The Puzzle

The proposition that large militaries undermine democracy, equity, and economic growth is widespread in Latin America. The most visible advocate of this position has been Nobel Peace Prize Laureate and former Costa Rican President Oscar Arias Saénchez. While not discarding the possibility that geography, history, and differences in class and power relations may explain part of the variance, Arias posits that a great part of Costa Rica's success in the post–WWII era stems directly from the dissolution of the military in 1948 and argues that other countries would also benefit greatly from demilitarization. For more than a decade, Arias has championed the vision that demilitarization is not a dependent variable, but rather a causal variable that can have significant positive developmental effects for Caribbean Basin nations individually and for the region as a whole.

When President Arias (1982–86) began exhorting his neighbors about the potential benefits of demilitarizing,the Cold War was raging. Arias was vilified by U.S. policy-makers and the very idea of another country going sans armée anywhere in Latin America was viewed by scholars as nothing more than a pipe dream. Even modest military reductions appeared unrealistic. In the past several years, the environment has radically changed. With the Cold War over and the neoliberal ideology of efficiency dominant, tremendous pressures are now being exerted on Caribbean Basin countries by U.S. embassies, J apanese aid officials, the International Monetary Fund, and local business elites to drastically downsize and in some cases eliminate armed forces. Panama recently disbanded its military, as has Haiti. Vigorous yet difficult efforts are under way in other Latin American countries to divert funds from the armed forces to debt reduction, education, and health services.

While Arias and others have used moral persuasion and case study evidence to argue that lower levels of militarization lead to development, more rigorous cross–sectional statistical analyses have not necessarily concurred. Indeed, the bulk of the quantitative research featuring inferential statistics shows a significant positive relationship between Third World militarization and economic growth and between militarization and equity/ social development. The vastly disparate results found by quantitative techniques and qualitative case studies represent a major intellectual puzzle. And this is a puzzle with important policy implications. With the end of the Cold War, domestic civil society groups and international actors are actively encouraging a reduction in military spending in many countries of the region even while a host of quantitative scholars arrive at the “somewhat awkward conclusion” that large militaries have a positive influence on economic growth and social development in Less Developed Countries (LDCs). The prime objective of this book is to solve this puzzle and firmly establish the relationship between militarization and development. The Research Design: Leverage and Dependent Variables The overarching argument—militarization has a negative effect on development in Latin America—is stated in very general,broad terms for two important reasons. First, this permits leverage maximization; we want to explain as much as possible with as little as possible (King, Keohane, and Verba 1994, 29). Second, research questions or propositions should permit clear tests of falsification. This very general argument allows more specific and observable implications of “development” that in turn generate testable hypotheses.

What is development? Mittelman defines development as the “increasing capacity to make rational use of natural and human resources for social ends” (1988, 22). While these types of definitions are useful in some contexts, they are also very difficult to operationalize. Huntington takes a different approach, defining five distinct goals of development: autonomy (from external forces), economic growth, democracy, equity, and stability. A great deal of research in the social sciences in recent decades has focused on one or more of these five goals. Early research under the modernization rubric deemed these goals complementary—all good things go together. Later research focused on the conflictual nature of these goals. Classic studies by Huntington (1968: modernization breeds instability), Kuznets (1955: economic growth leads to inequality), and O'Donnell (1973: economic growth leads to authoritarianism) are exemplars of the volumes of research dealing with the conflictual nature of the goals of development.

These goals of development are useful to my research for three reasons. First, they are much more easily operationalized than “development.” Second, they allow various empirical tests and opportunities for falsification. Third, having various components of development provides leverage; it is possible that we can explain a lot with only one causal variable. Having various dependent variables sacrifices parsimony and may make this project more complex than the typical study that focuses on one dependent variable. However, sacrificing parsimony for leverage is a profitable tradeoff; in this study I leverage the militarization causal variable.

I selected three of Huntington's five goals as separate dependent variables. There is little disagreement among mainstream scholars that democracy, equity, and economic growth are worthwhile goals. Stability and autonomy are another matter. I do not consider stability to necessarily be a long–term goal of development. Indeed, in my own research on the relationship between economic growth and equity, I conclude that a shake–up of the landed elite (instability) is strongly related with long–term economic growth and income equality. Many growth–and–equity success stories—Japan, South Korea, and Taiwan—experienced tremendous upheaval and instability that was a necessary requirement for a distribution of resources that set the stage for their long–term success (Bowman 1997).

Hirschman also points out that wars and instability are necessary conditions for peaceful resource redistribution (1963, 137). And Olson argues that shocks are needed to break up rent–seeking coalitions (1971, 1982). Collier and Collier weigh in very astutely on the inherent normative value of this variable: stability may be desirable under some circumstances, but under “other circumstances and from other normative perspectives, stability and the reduction of conflict may be seen as blocking needed change, whereas polarization may open new avenues for change” (1991, 10). Stability is not a goal of development under apartheid or when widespread exploitation occurs. Many Costa Ricans would argue that the 1948 Civil War was positive in the long term but that stability is inherently good now under a democratic regime. Democratic stability is a goal of development; authoritarian stability is not.

The relationship between militarization and autonomy is also not directly examined. “Autonomy” has been extraordinarily difficult to define and operationalize. I see little face validity in measures using United Nations voting behavior as an indicator of autonomy. In the globalized, interdependent world that emerged after World War II,one might also question the meaning and desirability of “autonomy from external forces.” No country, not even the United States, is autonomous from global economic forces.

This book directly examines three specific relationships stated in the following hypotheses:

Hypothesis 1: militarization has a negative effect on democracy

Hypothesis 2: militarization has a negative effect on equity

Hypothesis 3: militarization has a negative effect on economic growth

While I propose that militarization undermines development, I never claim that the militarization variable explains everything about Latin American development or even that it is the most important factor. Indeed, the multivariate analyses in Chapters 3 and 6 show that many variables affect democracy and development. Rather, I submit, and the evidence confirms, that the relationship is measurable, substantial, significant, and negative.

This research design uses both quantitative and qualitative methodologies, similar to Jeffery Paige's groundbreaking 1978 work. Good qualitative and good quantitative research designs share the same goals and logic (King, Keohane, and Verba 1994). However, the techniques are sufficiently different that each style has general strengths and weaknesses (Ragin 1987). Quantitative research generally lends itself to meeting higher standards of replicability, while qualitative research designs are generally more historically grounded. Quantitative designs are generally better for establishing relationships and their magnitudes; qualitative research is better for illuminating causal mechanisms and establishing sequence and agency.

For my research question, both methods can and will be used. Specifically, quantitative components were designed with the goal of establishing significant relationships and approximating the amount of the variance in the dependent variables explained by the explanatory variable. Qualitative sections are used to recheck the conclusions of the quantitative sections, flesh out causal mechanisms, document sequence and agency, and explore data and variables not amenable to quantification. The use of different methodologies and distinct sources and types of data to test a relationship is often referred to as triangulation.

Triangulation results from training distinct methodologies and datasets on the same empirical problem and then analyzing both the discrepant and similar results that obtain (Tarrow 1995, 473–74). King, Keohane, and Verba define triangulation as using the best methodology to analyze “data collected at different places, sources, times, levels of analysis, or perspectives, data that might be quantitative, or might involve intensive interviews or thick historical description” and concur with Tarrow that triangulation is the ideal in the quest of causal inference in the social sciences (1995, 479). If a large body of scholarship already exists, triangulation can marry existing research from one methodological school with original scholarly work on the same question but with different methods and different data sources. As quantitative research on Latin America is relatively underdeveloped and we have no existing studies on the topic in the time and space domain that interests me, I perform all of the analyses.

This research design, therefore, leads to both quantitative and comparative historical chapters. The two quantitative chapters are based primarily on eighteen Latin American cases, although Chapter 6 includes an examination of seventy–six LDCs. The comparative historical chapters focus on two countries, Costa Rica and Honduras. In the quantitative sections, my strong preference for studies of change over relatively long periods of time appears. Development is not a shortterm proposition and it may take many years for the impact of the causal variable to appear in the dependent variable. My research on the Kuznets inverted U–curve has awakened me to the problems of snapshot singleyear cross–sectional studies for demonstrations of causation (Bowman 1997). In this study, I focus on mid–to long–term trends of change in the dependent variables.

The Comparative Historical Case Selection

The logic of the comparative method has been described in a number of well–known works (Collier 1993; Lijphart 1975; King, Keohane, and Verba 1994). In a perfect research setting, one could easily gauge the effect of a causal variable on a dependent variable. One would start with two cases identical in every conceivable way. Then, the causal variable or treatment would be applied to one case and a placebo to the other. If the cases were identical in every aspect except the causal variable and an outcome or dependent variable differed greatly in the two cases, then we could conclude that there is some relationship between the causal and the dependent variables. Unfortunately, it is difficult if not impossible to create such an experiment in macro–social comparative research. The comparative method attempts to approximate the ideal. The approach is relatively straightforward. Cases should have variance in both the dependent variable and the causal or explanatory variable and little or no variance in other potential control variables that have been identified as having a potential impact on the dependent variable.

The selection process for this study began with the choice of Costa Rica as one of the cases,for it had the lowest level of militarization in the region. I then selected the other case. Since the five Central American Republics are often referred to as a single divided nation (Woodward 1985) and are all small countries with relatively small populations, I focused on El Salvador, Guatemala, Honduras, and Nicaragua. Honduras was selected as the second case, as it matches much more closely the case of Costa Rica on important control variables. Bananas and coffee are the two leading exports and the countries have very similar resource endowments. Neither case had the extensive forced labor conditions of El Salvador and Guatemala, and to a lesser degree of Nicaragua. Both are Mestizo countries and lack large marginalized ethnic groups that so strongly shape political conditions in Guatemala; although both Costa Rica and Honduras have had and continue to have small marginalized ethnic groups.

Both had until the middle of the twentieth century agricultural frontiers and an escape valve for those seeking land. Neither had a history of an enduring strong military institution as did Guatemala and El Salvador. Neither experienced a civil war during the era of the Cuban revolution,as did the other three countries. Both share long borders with Nicaragua and faced serious pressures from the United States during the Sandinista years. The Sandinista revolution and Contra wars make comparisons with Nicaragua problematic. Neither Costa Rica nor Honduras had a violently repressive oligarchy. And these countries had the two strongest union movements in Central America circa 1950.

As shown in Table 1.1, these two countries are remarkably similar on these important control variables—perhaps as similar as any pair of countries in the world—and yet are different in important ways besides militarization. These differences include the level of influence of transnational corporations in the economy, the development of the state, and most notably human capital as measured by literacy. The importance of literacy and human capital in Costa Rica cannot be overemphasized. Therefore, as with most two–case studies, the results are to a degree indeterminate. Yet, the real strength of comparative historical research, and the utility of combining it with inferential statistics, is in illuminating the causal mechanisms and showing agency and sequence in the relationships.

Chapter Previews

Chapter 2: Militarization: The Causal Variable, the Literature, and the Theory

Chapter 2 first precisely defines and conceptualizes the causal variable—militarization—and then presents an overview of the literature on the Third World military and identifies the important gap in the literature that this study fills. The theoretical foundations of much of the research on the military flows out of the groundbreaking work of Andreski (1954 [1968]), Huntington (1957), and Janowitz (1964). Three other important and distinct subgroups in the militarization literature are also introduced. First, I review the 1960s modernization literature that exhibits optimism for the modernizing effect of large Third-World armies. Second, I explore the quantitative literature that statistically assesses the relationship between militarization and various goals of development. And third, I incorporate the qualitative civil military literature that is focused on Latin America. A careful examination of these literatures results in a simple typology of Third World militaries. Then, using the explicit theoretical rationale of Andreski, Huntington, Janowitz, and others, I propose that militarization should have a negative impact on development in Latin America even if it has a positive impact in other regions in the developing world.

Chapter 2 ends with a discussion of the impact of militarization on state capacity and state priorities in Latin America. I propose that a combination of large militaries, internal military foci, and the international Cold War emphasis on anticommunism rearranged state priorities and undermined state capacity for development. The United States and many Latin American elites might have been in support of democracy and equity in theory, but during the Cold War anticommunism too often trumped democracy and conservative order trumped equity. Democracy and citizencentric development were not priorities and where state resources and attention were concentrated on internal order and security, state capacity to pursue the goals of development was undermined. Large militaries, particularly where enemies are fellow citizens, have serious opportunity costs.

Chapter 3: Taming the Tiger: A Quantitative Analysis of Militarization and Democracy in Latin America

The chapter argues that militarization in Latin America negatively affects three power relations that are important for democratization: class, transnational, and state. Two dimensions of democracy—average level of democracy over time and the lowest level of democracy during a multiyear period of time—are regressed against two indicators of militarization: military spending and the number of soldiers per thousand inhabitants. The results strongly support the hypothesis that militarization has a negative effect on democracy in Latin America. Finally, the chapter reveals that the negative impact of militarization on democracy began to diminish in the late 1970s, such that the negative effect is stronger in 1973–74 than in 1985–86.

Chapter 4: When Ballots Trump Bullets: Demilitarization and Democratic Consolidation in Costa Rica

Extant theories of Costa Rican democratic exceptionalism are presented and I argue that they are incomplete explanations of democratization. The ignored period 1948–58 must be studied to capture the causal mechanisms of democratic consolidation for this important case. In the 1948–58 period, the evidence is unambiguous that democracy was neither consolidated nor largely the result of elite pacts, Costa Rican culture, socioeconomic conditions, nor institutions. I illustrate how class power relations and transnational power dynamics threatened the Figueres regime both in 1948–49 and in 1953–58. The constitutional proscription of the military and the serious weakening of that institution restricted the options of the opposition who were united in their desire to destroy Figueres and who sought violent means to overthrow him. A preference for ballots over bullets only materializes after 1958 as the absence of a military as a deliberative actor affected the calculus of political elites.

Chapter 5: When Bullets Trump Ballots: Militarization and Democratic Collapse in Honduras

Honduras shared many similarities with Costa Rica during this period. In 1950,Honduras did not have a professional and institutionalized military. A moderate social democrat similar to Costa Rica's Figueres took power in 1957 and initiated an ambitious program of economic, political, and social reform. Like Figueres, President Villeda was opposed by the oligarchy, the banana companies, and the CIA. Villeda, like Figueres, was often smeared as a “communist.” One major difference between Costa Rica 1948–58 and Honduras 1954–63 was the very different trajectories of the armed forces. I detail the emergence of the Honduran military. In 1963, this new political actor ousted Villeda in a violent military coup. The chapter concludes that it would be difficult to imagine Figueres surviving with a strong Cold War military and that it would be easy to imagine Villeda surviving another day without U.S.–supported militarization of the country.

Finally, I argue that militarization was not the automatic destiny for Honduras. The decision to professionalize an army or follow the Costa Rican demilitarized example was a major public debate in the 1954–57 period. The debate continued during the Villeda presidency, and by the electoral campaign of 1963 was again a major national issue that was forcefully decided by the golpe de estado just days before the scheduled election.

Chapter 6: Guns Versus Butter: A Quantitative Analysis of Militarization and Material Development

This chapter has two purposes. The first is to assess the relationships between militarization and equity and between militarization and economic growth in Latin America. The second is to illuminate the importance of the boundedness of social explanation and illustrate how militarization has a different effect on equity and economic growth in Latin America than in sub–Saharan Africa or in the Middle East and East Asia. I first establish that food consumption is a useful indicator of equity. I assess the relationship between militarization and calorie consumption for eighteen Latin American countries for the 1964–89 period and find that militarization has a strong, highly signifcant,and negative effect on food consumption. The same methodology is then applied to the relationship between militarization and economic growth. The results strongly support the hypothesis that militarization has a negative effect on economic growth. This chapter also compares the militarization/food consumption and militarization/economic growth relationships in eighteen Latin American cases with a universal sample of seventy–six LDCs. I reveal how contextual differences in the militarization variable by region can lead to biased findings.

Chapter 7: Escaping the Lost Decade: Militarization and Economic Growth in Costa Rica and Honduras

Successful social and economic development requires sustained state intervention and two types of resources: economic resources and organizational resources. The decision to emphasize internal security and build up the armed forces has opportunity costs as scarce financial and organizational resources are consumed. This comparative historical chapter analyzes the trade–offs between militarization and economic development in the 1980s and 1990s, with foci on foreign–aid utilization and the tourism industry. I show how a greatly reduced military budget and a decision to pursue citizen–centric security rather than doctrines of national security in Costa Rica resulted in the state capacity and resources to transform the economy after the crisis of the early 1980s. In the rest of the Central American isthmus, the quest for internal security led to increased militarization during the Cold War. Large portions of the budget that could have gone to economic restructuring and infrastructure were consumed by the armed forces. Militarization also led to increased uncertainty, shortened time horizons, and a series of crises that thwarted the planning, implementation, and follow–through of the long–term policies that are necessary for economic development.

The book makes major contributions to both basic research and policy. A large body of statistical research on militarization and development is challenged and reassessed. The democratization scholarship for Costa Rica and Honduras is furthered and enhanced. In the Costa Rican case, the reassessment is considerable. This study also makes a significant positive contribution to the ongoing social science debate between the proponents of nomothetic findings and area studies. I show that area–specioc context is crucial for understanding development. Social science may progress further by seeking area–bounded explanation than universal truths. However, this study also shows that area–bounded explanations should both clearly identify those area–specioc manifestations that create context and carefully assess the claim that regional context alters the causal relationship.

Finally, the policy ramifications are immense. If militarization has an unambiguous negative effect on democracy, equity, and growth in Latin America, then civil society, politicians, business, and other domestic actors should make every effort to demilitarize the countries of the region. These efforts should be encouraged and supported by the United States, international financial institutions, and other international actors. Clinton and Bush administration initiatives to increase military sales to the region should be replaced with incentives to reduce the flow of arms into the hemisphere. Scarce resources should be diverted from the Latin American military and used to enhance democracy, reduce poverty, improve human capital, and lower debts. The era of the praetorian should end.

© 2002 Penn State University

History

           Latin American

Latin American Studies

           History

           Political Science

Political Science

           Comparative Politics

           Twentieth-century