The Impact of Economic Development on Democracy

Journal of Economic Perspectives—Volume 7, Number 3—Summer 1993—Pages 71–85 The Impact of Economic Development on Democracy Evelyne Huber, Dietrich Rueschemeyer, and John D. Stephens S ince the Second World War, two main research traditions have tackled the questions of which social and economic conditions most favor democracy: cross-national quantitative studies and comparative historical work. These two different methods have tended toward different theoretical positions, and more troublesome, arrived at contradictory results. One seminal work in the cross-national quantitative research program was Seymour Martin Lipset's (1959) essay on "Some social requisites of democracy: economic development and political legitimacy." These studies assembled a narrow range of aggregate data on development and democracy for many countries, converted them into standardized numerical values, and performed increasingly sophisticated statistical analyses of this material. Their theoretical interpretations were first inspired by modernization theory—a conception in which society, economy and polity are systematically interrelated, integrated by an overarching value consensus, and subject to increasing specialization and differentiation of social structures—while later research focused more on specific hypotheses and refrained from broader theoretical assumptions. Even though these studies used a variety of indicators for development and democracy and examined different samples of countries, they consistently arrived at one major result: the level of economic development correlated positively with • Evelyne Huber is Professor of Political Science, University of North Carolina, Chapel Hill, North Carolina. Dietrich Rueschemeyer is Professor of Sociology, Brown University, Providence, Rhode Island. John D. Stephens is Professor of Political Science and Sociology, University of North Carolina, Chapel Hill, North Carolina. 72 Journal of Economic Perspectives democracy. The interpretation of this result offered in these studies put primary emphasis on the spread of communication and education and the growth of the middle classes, all of which were supposed to lead to greater political interest and tolerance among growing numbers of citizens, thus creating the behavioral basis for democratic governance. This result was questioned by a number of comparative historical studies. The most eminent example of this alternate line of research was Barrington Moore's The Social Origins of Dictatorship and Democracy (1966). Moore's work, as well as other historical studies of the relationship between capitalist development and democracy from Max Weber (1906) to Guillermo O'Donnell (1973), have come to skeptical conclusions about the chances of democracy as capitalist economic development spread around the globe. These studies were built on the intensive analysis of a few cases and informed by a conflict-oriented political economy approach—an approach which places the relationship between economic and political power at the center of analysis, puts less store by the systemic integration of societies, and sees social change as driven by conflicting interests of social actors. As Moore (1966, p. 5) expressed the essence of their conclusions, "the route that ended up in capitalist democracy ... was itself a part of history that almost certainly will not be repeated." We took this impasse in research on the impact of development on democracy as the point of departure for our own work, and we hope that our analysis has broken the impasse. Methods, Theory, and Major Results Any account of the social and economic conditions of democracy must come to terms with the central finding of the cross-national statistical research: a sturdy (though not perfect) association between economic development and democracy. But these correlations do not validate the theoretical accounts that have often been associated with them, in particular modernization theory. Nor do cross-sectional correlations allow us to make adequate inferences about causal sequence. Similar outcomes might be produced by a variety of factors and causal sequences. To tackle these questions of causation, we adopted a strategy of analytic induction based on comparative historical research. This strategy is a case-based method of study, which builds on a theoretical framework that takes past research into account, and then proceeds by analyzing successive individual histories. In this way, this method gains information on historical sequence and can do justice to the particular historical context of each factor analyzed. Each case may modify both the specific hypotheses used in earlier analyses and the broader theoretical framework. The result is a range of cases interpreted by a single set of theoretical propositions and a progressively modified theory that is consistent with the cases studied. Evelyne Huber, Dietrich Rueschemeyer, and John D. Stephens 73 From this perspective, describing our methods and theory involves specifying which cases we examined; the working definition of democracy we used; specifying a dynamic by which democracy emerges; and then finally, developing insights into how economic development affects that dynamic. Case-based research typically encounters the problem of working with too few cases and too many variables. We sought to reduce this problem by stretching the possibilities of comparative historical analysis to the limit and examining a large number of countries in three groups: the currently advanced capitalist countries, South America, and the Central American-Caribbean basin. These countries include most democracies in the world today and most historical experiments with democracy; they also represent a larger number of cases than are used by quite a few cross-national statistical studies. What is the theoretical framework we used to analyze these histories? We begin by defining democracy as having three features: regular free and fair elections of representatives on the basis of universal suffrage; responsibility of the state apparatus to the elected representatives; and guarantees of freedom of expression and association. This definition, though relatively standard in the literature on the subject, gives greater weight than common usage to the universality of suffrage, in particular to the extent in which the right to vote cuts across boundaries of region, religion, ethnicity, and—above all—social class.1 Thus, neither the Greek city states, which excluded slaves and residents without citizenship from suffrage, nor mid-nineteenth century Great Britain, which denied the vote to the vast majority of citizens, were democracies by this definition, even though in both cases public policy was openly discussed and voted on. Democracy must in our view mean at a minimum a significant share of the many in political decision-making. However, while our comparative historical work emphasizes the attainment of full democracy, it also extends to the building blocks for full democracy, such as responsible government with restricted suffrage, and thus its validity does not stand and fall with this emphasis on universal suffrage. Our most basic assumption is that democracy is a matter of power and power sharing. This fundamental premise led us to focus on three power clusters as primarily relevant for the chances of democracy: (1) the balance of class power as the most important aspect of the balance of power in civil society; (2) the nature of the state and state-society relations, or the balance of power between state and civil society; and (3) transnational structures of power, or the international economy and system of states, as they shape the first two balances and constrain political decision-making. For a society to become democratic, the power balance in civil society has to shift. Civil society is the public sphere distinguished from the state, the economy and the web of family and kin relations. It comprises all social groups, 1 We choose universal male suffrage as a criterion because the extension of female suffrage is driven by a different dynamic and thus would require a separate analysis of considerable length. 74 Journal of Economic Perspectives associations and institutions that are not strictly production-related, nor governmental or familial in character. Since the major power resource of the many is collective organization, their chance to organize in associations, unions and parties gains critical significance. Among social classes we focus in the first place on dominant and subordinate positions—on landlords and peasants in the agrarian sector and on major capital owners (classically labeled the bourgeoisie) and workers in the industrial sector. Independent small and medium farmers, craftsmen, merchants, and the new white collar employees—often collectively referred to as the middle-classes —stand in between; but before the attainment of democracy they, too, were in most countries excluded from political participation. Democratic participation in political decision-making will develop and be sustained only if the economic and cultural power of dominant groups is counterbalanced in civil society by the organizational power of subordinate classes. The structure of the state and state-society relations are also clearly relevant for the chances in democracy. The state needs to be strong and autonomous enough to ensure the rule of law and avoid being the captive of the interests of dominant groups; the state's authority to make binding decisions in a territory and the state's monopoly of coercion must be settled. The vote does not rule where it competes with the gun. However, the power of the state needs to be counterbalanced by the organizational strength of the civil society to make democracy possible; the state must not be so strong and autonpmous from all social forces as to overpower civil society and rule without accountability. Thus, centralized state control over the economy and the presence of a large military and police apparatus are conditions inimical to a favorable power balance between state and civil society. The relation of the state to religious organization is another critical factor. Civil society gains strength from non-established religious movements, while an alliance of "crown and altar" strengthens the hand of the state. The third power cluster involves international power relations. Aside from the impact of war (typically creating a need for mass support and discrediting ruling groups in case of defeat), we look especially into the role of economic and geopolitical dependence. Based on cross-national statistical results and influenced by the more historical work of the new dependency theorists (Cardoso and Faletto, 1979), we expected dependency to be an important factor but one without a clear-cut, unequivocal effect. The three power clusters—relative class power, the role of the state, and the impact of transnational power structures—are closely interrelated. For instance, economic dependency can have long term effects on the structures of class; war and geopolitical factors can strengthen the role of the security forces within the state; and the results of power relations in civil society are crucially affected by differential access to the state apparatus. Our central thesis, and indeed our most basic finding, can now be stated in stark fashion: Capitalist development is related to democracy because it shifts The Impact of Economic Development on Democracy 75 the balance of class power, because it weakens the power of the landlord class and strengthens subordinate classes. The working and the middle classes—unlike other subordinate classes in history—gain an unprecedented capacity for self-organization due to such developments as urbanization, factory production, and new forms of communication and transportation. This thesis negates other explanations. The primary link between capitalist development and democracy is not found in an expansion of the middle classes. Nor can the relationship be explained by the argument that more complex societies require a differentiated and flexible form of government, as modernization theory suggested. And finally democracy is not the creation of the bourgeoisie, the new dominant class of capital owners, as was claimed by both liberal and Marxist political theory. The bourgeoisie made important contributions to the move towards democracy by insisting on its share in political power in the form of parliamentary control of the state, but the bourgeoisie was also hostile to further democratization when its interests seemed threatened. In fact, one of the more important findings of our comparative research, which we did not fully anticipate, is that—especially in Latin America—the economically dominant classes accepted democracy only where their political interests were effectively protected by large parties of a conservative or non-ideological character. It is also important to note that the bourgeoisie often comes around to support democracy once it turns out that its interests can be protected within the system. Having stated the thesis bluntly, some warnings to the reader are in order. We are not arguing that the correlation between level of development and degree of democratization is unilinear or automatic. And we certainly are not arguing that class is all that matters. Although we consider the shift in the balance of class power to be the most important factor accounting for the positive correlation between development and democracy, our analysis leaves ample room for the other two clusters of power and for complex interactions among them. Perhaps the biggest complicating factor, given the centrality of the balance of class power in our overall interpretation of the association of development and democracy, is that class interests are not ahistorical givens; they are historically constructed by movements, organizations and leaderships that act in some particular environment of influences and oppositions, possible alliances and enmities. Once set, these constructions often persist beyond the constellation of origin. "Working class interests" are quite different when determined by social democratic parties than they would be as determined by Catholic or by Leninist parties. The fact that class interests are historically constructed has crucial consequences for the analysis. It raises interclass relations to critical importance. One class may exercise hegemonic influence over another, and this will affect the alliance options among classes. The interests actually pursued by peasants and even by urban middle classes are often profoundly shaped by landlords, the 76 Journal of Economic Perspectives bourgeoisie, and the state as well as state-affiliated churches. The alliance developments at the top—among landlords, bourgeoisie, and the state—can be decisive for the alliance options of other classes. This is of critical importance for the chances of democracy because the working class, even the European working class, was too weak on its own to succeed in the final push toward democracy with universal suffrage. Democratic Transition and Breakdown in Europe and South America In 1870, only one European country, Switzerland, was a democracy. Many countries frequently thought to be democratic at this time such as Britain, Netherlands, and Belgium, had parliamentary government and competitive party systems, but the electorate was limited by income or property qualifications. By contrast, by 1920, almost all Western European countries were fully democratic. This period of transition to democracy in Europe was also marked by the arrival of the organized working class. The change in the underlying class structure as indicated by labor force figures is significant enough: between 1870 and 1910, the non-agricultural workforce grew by one-third to one-half, eventually reaching an average of 61 percent of the total workforce in the 13 European countries we studied. The change at the level of class formation and class organization was even more significant: in 1870, in no country were the socialists a significant mass-based party, and the trade unions organized a minuscule proportion of the labor force. By the eve of World War I, the major socialist and labor parties garnered an average of 26 percent of the vote (despite suffrage restrictions in a number of countries) and the trade unions organized an average of 11 percent of the non-agricultural labor force. In the immediate postwar elections, the socialists' electoral share increased to an average of 32 percent, while trade union organizations grew spectacularly, increasing two and a half times. The organized working class was also the most consistently pro-democratic force in the period under consideration: at the onset of World War I, European labor movements had converged on an ideology which placed the achievement of universal suffrage and parliamentary government at the center of their program (Zolberg, 1986). Though the working class was the main agent of democracy in Europe, it needed allies. It found them in the urban middle classes and the independent small farming population. Without these groups, the working class was too weak to press through full democracy. Indeed, in Switzerland and Norway, two countries which (like the north and west of the United States during the Jacksonian period) might be termed agrarian democracies, these groups were more important in the struggle for democracy than the working class. Evelyne Huber, Dietrich Rueschemeyer, and John D. Stephens 77 However, unlike the working class, both the urban middle classes and small farmers were not consistently pro-democratic: In some countries, they were ambivalent about the introduction to democracy and, in the interwar period, they provided mass support for fascism and other authoritarian movements which destroyed the new democratic regimes. What was the role of the bourgeoisie, the class of the major owners of capital? In only three of the 13 European countries studied—France, Switzerland, and Britain—did any significant segment of the bourgeoisie play a leading role in promoting full democracy. Significantly, in all three of these cases, the bourgeoisie did not face a working class politically organized by socialist parties at the time of democratic transition; in ten of the other eleven countries, it did face such an opponent. Fear of challenges to property rights certainly played an important role in the reticence of propertied upper classes to support political inclusion of the working class. As Barrington Moore has argued, the existence of a powerful class of landlords dependent on a large supply of cheap labor were associated with significant problems for democracy: In four of the five Western European countries in which large landholders played a significant political role towards the end of the nineteenth century—Germany, Austria-Hungary, Spain and Italy—democratic regimes collapsed in the interwar period. In each of these countries, the landed upper classes, in coalition with the state and the bourgeoisie, were crucially implicated in the weakness of the push toward democracy outside of the working class movement before World War I and in the events that led to the demise of democracy in the interwar period. Does Latin America show similar developments? Patterns for large landholding and the existence of a powerful class of landlords with a need for a large cheap labor force also posed significant problems for democracy in South America. Breakthroughs to full democracy before the 1970s, even if temporary, were only possible where the large landowners were primarily engaged in ranching and thus had lower labor needs (Argentina and Uruguay), or where their economic power was undermined or counterbalanced by the presence of a strong mining export sector (Venezuela and Bolivia). Like its counterpart in Europe, the bourgeoisie was not a promoter of full democracy in South America. As in Europe, the forces pushing for democracy were the organized segments of the subordinate classes, but the leadership roles were reversed. In South America the middle classes were the driving force, but they mainly promoted their own inclusion and thus often accepted restricted forms of democracy. For full democracy to be installed, the middle classes had to be dependent on working class support in their push for democracy, and they had to receive support from a working class which had some measure of strength. Peron's Argentina (1946–55) provides a dramatic illustration that the working class was not invariably pro-democratic, and could be attracted to 78 Journal of Economic Perspectives support authoritarian rulers, if these rulers were the first ones to include the working class on a large scale and to promise to satisfy its material demands in the meaningful way. The political history of 20th century Latin America is characterized by numerous breakthroughs to restricted or full democracies, then followed by breakdowns of democracy. Essentially, the economically dominant classes tolerated democracy only as long as what they perceived to be their vital interests were protected. Where the capacity of the state or political parties to channel and contain militant action of subordinate classes declined, economic elites turned to the military in search of allies to replace the democratic with authoritarian regimes. State structure and state-society relations were inimical to democratization in Latin America in several ways. State consolidation had to be achieved in a much shorter period and later relative to economic development than in Europe, and without a consolidated state there could be no democracy. The independence wars (1810–25) and later wars over borders (for example, the War of the Pacific, Peru and Bolivia against Chile, 1879–83) left the societies with a strong legacy of militarism. Furthermore, rather than industrialization leading to urbanization and an export economy, as happened in Europe, Latin America saw the growth of the export economy and of urbanization preceding industrialization. This gave the state additional power, and in many cases led to attempts by the state to preempt an independent organization of the emerging industrial working class. Finally, in the post-World War II period the state's coercive capacity grew stronger, in part due to U.S. military assistance, as did the state's capacity to mobilize economic resources independent of domestic economically dominant classes. This gave the state greater autonomy from civil society and generated a new form of anti-democratic regimes, called bureaucratic-authoritarian regimes, in which the military as an institution exercised power and engaged in large-scale repression of labor movements and reformist political parties (as in Brazil after 1964, Chile after 1973, and Argentina after 1976). These developments also involve the impact of the third power cluster, the ways in which the international economy and system of states were important for the trajectory of democratization in South America. The crucial role of the state as intermediary to international markets for goods, capital, and technology afforded the state significant autonomy from civil society. This autonomy was reinforced through external support for the security forces. The position of South American countries in the world economy as late and dependent developers, with imported technology, resulted in small industrial working classes compared to Europe at similar levels of economic development, and thus in class structures inimical to democratization. Economic dependence further meant high vulnerability to fluctuations in world markets, and the resulting economic instability made stabilization and legitimization of regimes difficult, whether those regimes were authoritarian or democratic. The Impact of Economic Development on Democracy 79 Central America and the West Indies The West Indies and the Central American countries share a number of socioeconomic characteristics that have been shown to be inimical to democracy.2 Their economies were traditionally plantation economies, with some mining and industrialization—and tourism in the West Indies—superimposed in the post-World War II period. The corresponding societies were traditionally very hierarchical and their economies highly dependent on foreign trade and foreign investment. Low economic development, extremely high dependence, high inequality, a small working class, and rapid social change all have been found to be unfavorable for the installation and consolidation of democratic regimes, and they all characterized Caribbean Basin countries in the post-World War II period. Not surprisingly, then, all but two of the Spanish-speaking countries in the Caribbean basin were ruled by authoritarian regimes in the 1960s and 1970s, and during the 1970s, economic elites and the military establishment resorted to increasingly violent repression of both revolutionary movements and democratic reformist forces. The exceptions were Costa Rica and, from 1978 on, the Dominican Republic. In contrast, all but two of the English-speaking Caribbean countries had democratic regimes from the time of their independence in the 1960s throughout the 1970s.3 How do we explain these contrasting political developments? The antecedents of the contrasts in the 1960s and 1970s lie in the developments of the 1930s. The Depression brought great disruptions to the extremely export-dependent societies in the region. In response to decreasing real wages and increasing unemployment, attempts at labor organization and labor protests emerged in virtually all countries throughout this region. The reactions of the economic elites to these protests and organizing attempts were universally negative, but the reaction of the state varied widely. British colonialism was important here, because it was an alternative to the Central American pattern of landlord or military control of the state, and thus an alternative to the use of the coercive forces of the state to repress both the protests and the emerging labor unions and allied political parties. Consequently, the '30s marked the beginning of organized political life, and opened the way for the subsequent consolidation of civil society in the West Indies, whereas in Central America they solidified the pattern of the primacy of the coercive apparatus of the state and of state control over the repression of civil society, exercised either by 2 The term "West Indies" refers to the English-speaking Caribbean countries (the larger ones are Jamaica, Barbados, Trinidad, and Guyana); the term "Caribbean" includes them along with Spanish, French, and Dutch-speaking countries; the term "Caribbean Basin" refers to Caribbean and Central American countries. 3 Racially polarized Guyana and tiny Grenada were the exceptions, but analysis of these two countries would take us too far afield here; the interested reader should see Rueschemeyer, Stephens and Stephens (1992, pp. 251–58). 80 Journal of Economic Perspectives landowner-military coalitions or the military alone. Costa Rica, the deviant case in Central America, resembled the West Indies insofar as the large landowners were not in firm control of the state apparatus (though for different reasons, as we shall see in a moment) and consequently unions and political parties were allowed to consolidate their organizations. To explain why the economic elites and the military controlled the state in Central America countries and the Dominican Republic in the '30s, and why they used this control to squash emerging social forces, two interdependent sets of factors are crucial: the economic strength of the large landowners and their relationship to the state; and the extent of U.S. economic presence and direct political intervention. To underline the point that the role of the state and foreign forces has to be taken into consideration, along with the strength of the landed class, one can distinguish two basic types of class-state constellations in Central America in the first three decades of this century (setting aside Costa Rican exceptionalism). In one type, exemplified by El Salvador and Guatemala, the large landowners were very prosperous, having established commercial and financial holdings as well (El Salvador), or having merged with originally merchant elites (Guatemala). They formed an oligarchy in the true sense of the word, controlling the state directly or via military officers. In the other type, exemplified by Honduras, Nicaragua and the Dominican Republic, the landowners were not as prosperous because historically the territories had been sparsely populated, had suffered many military invasions and, particularly in Nicaragua, had seen fierce fights among different factions of landowners. Also, foreign capital had a strong presence, dominating the financial system and, in the case of Honduras, the crucial export sector (bananas). Thus, the landowners as a group were not in control of the state; rather, military strongmen exercised state power. Moreover, these three countries were the object of repeated U.S. military intervention and occupation in the first third of this century, which further impeded the establishment of civilian control over the state apparatus. In sharp contrast, to the situation in the Spanish-speaking Caribbean, landlords in the British West Indies did not control the state. Britain did; and British rule acted as a restraining influence. Planters called for the use of troops to put down labor unrest, for public flogging to set an example, and so on, but the colonial authorities allowed unions based in the working class and nationalist political parties based in the middle classes to emerge and to develop into the driving forces for internal democratization and decolonization. State action responded to the balance of class power in England, not in the Caribbean. A dramatic illustration of the tremendous contrast in state reactions to labor unrest in the '30s can be seen, on the one hand, by the death of 10,000 to 30,000 Salvadorans (depending on the estimate) who were slaughtered by the coercive forces in 1932, and, on the other, by the comparatively few labor leaders who were imprisoned, mostly for a short time, in Jamaica. This difference in reaction was not rooted in a difference of landlord interests. West Indian landlords favored repressive action at that time also. Evelyne Huber, Dietrich Rueschemeyer, and John D. Stephens 81 If one turns back in history to a period when the West Indian landed upper classes did have much more control of the state—for example, to the period before the introduction of crown colony rule in the late 19th century—one can get an indication of what the planters might have done if they had controlled state power later. As one example, consider the Morant Bay rebellion in Jamaica in 1865. Police fired on a protesting crowd and killed seven people. This set off a riot in which 14 people were killed. The plantocracy reacted hysterically and had 439 people hanged, 600 flogged, and 1000 black peasants' houses burned. The relatively moderate reaction of Britain should not be taken to imply that the British Colonial Office was a leading promoter of democracy and decolonization. As the eminent Caribbeanist Gordon Lewis (1968, p. 108) puts it: "Colonial Office policy... was to grant minuscule reforms at the last moment, discriminating between different territories, and seeking every way to delay the inevitable; and progress in any case, was the result of militant progressive forces in each colony, extracted from London through protest and agitation." Those who did press the process forward were alliances of organized labor and politically organized middle classes. The speed and durability of this progress depended on how strong and cohesive this coalition was. This middle class-working class coalition bears some resemblance to the European democratic political coalitions; yet two contrasts stand out. First, the West Indian movements were much more middle-class dominated and, second, the West Indian middle-classes showed less equivocation about democracy. Ethnicity certainly played a role here: the ability of the brown middle classes to displace the white upper classes from political leadership depended on the enfranchisement of the black lower classes. At any rate, that is what the history of suffrage expansion in the region indicates. Not only the plantocracy, but also capitalists in commerce and finance opposed the democratic and independence movement. This is indicated by alignments of the political leaders of business in the immediate post-World War II decades, and it is confirmed by a systematic study of West Indian leaders which showed that 94 percent of the West Indian businessmen opposed independence (Moskos, 1967, p. 42) Costa Rica deviated from the Central American pattern in that the dominant class had its main economic base in control over commerce and finance, not land and labor, and the expansion of the coffee economy in the 19th and early 20th century into frontier areas produced a relatively autonomous agrarian middle class, a significant urban artisanal middle class, and a class of small farmers. Thus, in the critical juncture of the 1930s, there was no hegemonic class of landowners exercising firm control of the state, and a sphere for the organization of civil society also opened up. A second surge of militance arose in Central America and, to a lesser extent, in the West Indies in the 1960s and 1970s. In the West Indies, the strategy of "industrialization by invitation" (that is, based on foreign 82 Journal of Economic Perspectives investment), following in part the economic development prescriptions of Sir Arthur Lewis, and particularly the investment in bauxite, oil, and tourism in the 1950s and 1960s, led to a growth of the working class and the urban middle classes, but also to increasing unemployment and inequality. In Central America, similar developments occurred with the growth that was spurred by the Central American Common Market in the 1960s. There, the growing disparities were even more pronounced, as the spread of commercial agriculture caused great dislocations in the agrarian sector. In both regions, new unions, middle-class professional associations, and political parties were formed to promote the interests of these new groups, and existing organizations expanded and put forward demands for reforms. In Central America, most states responded again with repression. The repressive apparatus had been further developed with U.S. aid in the postWorld War II period, particularly in response to the Cuban revolution, and it had become increasingly autonomous from civilian elites by the 1970s. This is not to say that the economically dominant classes favored democratization over repression; on the contrary, particularly in El Salvador and Guatemala, they heartily supported political and economic exclusion of the lower classes and death squad activities. Repression in turn fueled revolutionary challenges, leading to the overthrow of the personalist (or familial) dictatorship in Nicaragua and to prolonged civil strife in El Salvador and Guatemala. In Honduras, Panama, and the Dominican Republic after 1965, the authoritarian regimes were less repressive and the popular challenges less violent. Nevertheless, the weakness of civil society and U.S. support for the military contributed to the continued failure of pressures for democratization. A dramatic, but temporary, turn in this foreign influence towards stronger support for democratization came with the pressure exerted by the Carter administration on the military in the Dominican Republic to respect the 1978 elections. Costa Rica was again critically different. After the civil war of 1948, which was not an insurgency rooted in excluded subordinate classes but rather a challenge of counter-elites to a government attempting to stay in power through election fraud, the military was dismantled. Subsequently, civil society continued to grow in strength, as did political parties. When the U.S. exerted pressures for militarization in the 1980s, in the context of its efforts to put pressure on the Sandinista government in Nicaragua, the political system and civilian political elites were consolidated enough to be able to resist these pressures and thus keep this potential threat to democracy from emerging. In the West Indies, where control of the state passed to local political elites with independence in the '60s, unions and parties had grown stronger and were for the most part able to integrate the newly emerging social forces and channel their militancy through democratic institutions. Moreover, the repressive apparatus remained small, and the economically dominant classes had learned to work through the parties to defend their interests and thus had come to accept democracy. The Impact of Economic Development on Democracy 83 In sum, the growth of democracy cannot be read off from the economic development and its effect on the development of the class structure alone. Both Central American and Caribbean countries started from roughly similar levels of development, similar economic structures, and similar world market niches. What emerges as critical in this comparison of the emergence of democracy in the West Indies and its absence in Central America is the nature of state-class relations, especially the critical contribution of British colonialism. Similarly, the development of a strong military state or its absence cannot be explained by internal factors alone. The third power cluster, transnational structures of power, must be brought in to complete the explanation. The United States' economic and geopolitical interests, along with political alignments within the United States, led it to support the build-up of the coercive apparatus of the Central American states. In Britain, by contrast, the colonies were increasingly viewed as an expense which a declining power could ill afford and which the social democratic forces, which were far stronger in Britain, no longer desired. This analysis of the West Indies and Central America leads to a reinterpretation of our evidence on Europe, as one would expect in this strategy of analytic induction. In the discussion of Europe, we attributed the authoritarian trajectory of four of the countries—Germany, Italy, Austria-Hungary, and Spain—to the strength of the landed upper classes in those countries. The West Indian cases indicate that this strength had to be complemented with a structure of state power which was open to strong landlord influence and could be turned to coercive purposes on a national scale. Such a state had been created across the European continent in the course of the centuries-long consolidation of the European states, largely as a result of warfare between the states. As Tilly (1975, pp. 40–44) points out, the coalition of large landlords with the central state was a militarily strong one and was frequently victorious in these centuries of war. Moreover, the state-building process in these countries generally resulted in the landed upper classes having a strong foothold in the military, which made recourse to authoritarian politics more attractive since landlords and their allies could rely on the military to exercise the coercion necessary to maintain or install authoritarian rule. Questions, Method, and Applications Our program of comparative historical research confirmed the conclusion of the cross-national statistical analyses of the correlates of political democracy: The level of economic development is causally related to the development of political democracy. However, the underlying reason for the connection, in our view, is that capitalist development transforms the class structure, enlarging the working and middle classes and facilitating their self-organization, thus making it more difficult for elites to exclude them politically. Simultaneously, 84 Journal of Economic Perspectives development weakens the landed upper class, democracy's most consistent opponent. The development of the class structure hardly accounts for all national differences in democratic development, as the contrasting political development in the Spanish-speaking Central American countries and the English-speaking Caribbean islands demonstrates, but it is of central importance. Some readers may be familiar with the argument that the bourgeoisie played an important role as the agents of democratic reform, and thus may be surprised at how little weight we give this factor. Surely, leading businessmen in contemporary advanced capitalist countries are rightly regarded as supporters of democracy. However, most of their predecessors in 19th century Europe and 20th century Latin American were not, because they feared that extending suffrage to workers would represent a threat to their material interests. As democracy was established during the 20th century and these fears proved to be exaggerated, the bourgeoisies of advanced capitalist societies gradually came to accept and then strongly to support democratic institutions. A similar process occurred in the West Indies in the post-World War II period, and one can hope that contemporary South America is experiencing the same phenomenon. A brief reflection on method complements the substantive review. Our analysis offers a causal theory about the long-term conditions favoring and obstructing democratization. More than this brief review could show, our research incorporates the results of cross-national statistical work, while it is itself based on extensive comparative historical analysis. We consider this combination as the best way of dealing with the thorny problems of macrosocial analysis, in particular the problems created by the limited number of cases, the importance of sequence analysis for causal arguments, and the possibility of different causal paths leading to similar outcomes. We conclude with a few observations about the implications of our analysis for the future of democracy in the contemporary Third World. Across the less developed countries, most U.S. intervention since World War II (at least) was primarily motivated by geopolitical competition with the former Soviet Union, rather than by direct defense of economic interests. Concerns for national security were invariably invoked to justify support for authoritarian regimes. The end of the Cold War both alleviates these concerns and greatly weakens them as a basis for foreign policy towards Latin America. The discrediting of the Soviet model has also dealt a mortal blow to Leninist socialism as a model for opposition movements and a basis for legitimization of authoritarian regimes in the Third World. These factors at least open the door for a more unambiguously pro-democratic policy towards the Third World on the part of the United States and the other developed countries, all of which are now democracies. But other factors lead to less optimism about the chances for democracy in the Third World. In the case of both workers and businessmen, our analysis shows that their political posture toward democratic institutions was motivated in no small part by their perception of how democracy would affect their Evelyne Huber, Dietrich Rueschemeyer, and John D. Stephens 85 material interests. On this account, one can say that the current economic problems in the Third World, economic stagnation and the crushing debt, are also a problem for democracy. There is no doubt that rapid economic growth, or a growing economic pie, facilitates compromise between capital and labor and that, conversely, slow growth makes it almost impossible to satisfy both parties. Under such conditions, demands for mere economic betterment on the part of workers become a threat to business. The analysis leads us to expect some countries within the Third World to have better prospects for democratization than others. Most obviously, the prospects are brighter for those countries at higher levels of economic development. However, as our analysis made clear, it is not the mere rise in per capita income (created, for example, by mineral wealth) that is of greatest importance, but rather the changes in the class and social structure caused by industrialization and urbanization which are most consequential for democracy. In addition, the analysis of agrarian class relations leads us to the conclusion that democratic prospects are much better in Third World countries without a significant group of large landholders and with a significant agrarian middle class. • This article presents a summary of the central arguments made in our book Capitalist Development and Democracy (Rueschemeyer, Stephens and Stephens, 1992). The change in the order in which our names appear is due exclusively to the name change of one of the co-authors; the order remains alphabetical and does not represent any statement about our relative contributions. We expressed in the book our indebtedness to a large number of colleagues. Dietrich Rueschemeyer wishes to acknowledge in addition with deep gratitude the hospitality of the Swedish Collegium for Advanced Study in the Social Sciences, Uppsala, and the comments made by colleagues in a three-day colloquium on "Democracy and Societal Developments" held at the Collegium in October 1992. 86 Journal of Economic Perspectives References Cardoso, Fernando Enrique, and Enzo Faletoo, Dependency and Development in Latin America. Berkeley: University of California Press, 1979. Lewis, Gordon, The Growth of the Modern West Indies. New York: Modern Reader, 1968. Lipset, Seymour Martin, "Some Social Requisites of Democracy: Economic Development and Political Legitimacy," American Political Science Review, 1959, 53, 69–105. Reprinted in Lipset, S. M., Political Man, expanded edition. Baltimore: Johns Hopkins University Press, 1980, 27–63. Moore, Harrington, Jr., The Social Origins of Dictatorship and Democracy. Boston: Beacon Press, 1966. Moskos, Charles, The Sociology of Political Independence. Cambridge: Schenkman, 1967. O'Donnell, Guillermo, Modernization and Bureaucratic Authoritarianism. Berkeley: Uni- versity of California at Berkeley, Institute of International Studies, 1973. Rueschemeyer, Dietrich, Evelyne Huber Stephens, and John D. Stephens, Capitalist Development and Democracy. Cambridge. Polity Press, and Chicago: Chicago University Press, 1992. Tilly, Charles, "Reflections on the History of European State-making." In Tilly, Charles ed., The Formation of National States in Western Europe. Princeton: Princeton University Press, 1975, 3–83. Weber, Max, "Zur Lage der bürgerlichen Demokratie in Russland," Archiv fur Sozialwissenschaft und Socialpolitik, 1906, N.S. 22 Zolberg, Aristide R., "How Many Exceptionalisms?" In Katznelson, Ira, and Aristide Zolberg, eds., Working Class Formation: Nineteenth Century Patterns in Western Europe and the United States. Princeton: Princeton University Press, 1986, 397–455.