Schmitter & Karl What Democracy is and is not

What Democracy Is. . . and Is Not Philippe C. Schmitter, Terry Lynn Karl Journal of Democracy, Volume 2, Number 3, Summer 1991, pp. 75-88 (Article) Published by The Johns Hopkins University Press DOI: 10.1353/jod.1991.0033 For additional information about this article http://muse.jhu.edu/journals/jod/summary/v002/2.3schmitter.html Access provided by University of California, San Diego (9 Jan 2014 01:21 GMT) WHAT ... DEMOCRACY A N D IS N O T IS Philippe C. Schmitter & Terry Lynn Karl Philippe C. Schmitter is professor of political science and director of the Center for European Studies at Stanford University. Terry Lynn Karl is associate professor of political science and director of the Center for Latin American Studies at the same institution. The original, longer version of this essay was written at the request of the United States Agency for International Development, which is not responsible for its content. For some time, the word democracy has been circulating as a debased currency in the political marketplace. Politicians with a wide range of convictions and practices strove to appropriate the label and attach it to their actions. Scholars, conversely, hesitated to use it--without adding qualifying adjectives--because of the ambiguity that surrounds it. The distinguished American political theorist Robert Dahl even tried to introduce a new term, "polyarchy," in its stead in the (vain) hope of gaining a greater measure of conceptual precision. But for better or worse, we are "stuck" with democracy as the catchword of contemporary political discourse. It is the word that resonates in people's minds and springs from their lips as they struggle for freedom and a better way of life; it is the word whose meaning we must discern if it is to be of any use in guiding political analysis and practice. The wave of transitions away from autocratic rule that began with Portugal's "Revolution of the Carnations" in 1974 and seems to have crested with the collapse of communist regimes across Eastern Europe in 1989 has produced a welcome convergence towards a common definition of democracy. ~ Everywhere there has been a silent abandonment of dubious adjectives like "popular," "guided," "bourgeois," and "formal" to modify "democracy." At the same time, a remarkable consensus has emerged concerning the minimal conditions that polities must meet in order to merit the prestigious appellation of "democratic." Moreover, a number of international organizations now monitor how well Journal of Democracy Vol.2,No.3 Summer 1991 Journal of Democracy 76 these standards are met; indeed, some countries even consider them when formulating foreign policy.2 What Democracy Is Let us begin by broadly defining democracy and the generic concepts that distinguish it as a unique system for organizing relations between rulers and the ruled. We will then briefly review procedures, the rules and arrangements that are needed if democracy is to endure. Finally, we will discuss two operative principles that make democracy work. They are not expressly included among the generic concepts or formal procedures, but the prospect for democracy is grim if their underlying conditioning effects are not present. One of the major themes of this essay is that democracy does not consist of a single unique set of institutions. There are many types of democracy, and their diverse practices produce a similarly varied set of effects. The specific form democracy takes is contingent upon a country's socioeconomic conditions as well as its entrenched state structures and policy practices. Modern political democracy is a system of governance in which rulers are held accountable for their actions in the public realm by citizens, acting indirectly through the competition and cooperation of their elected representatives. 3 A regime or system of governance is an ensemble of patterns that determines the methods of access to the principal public offices; the characteristics of the actors admitted to or excluded from such access; the strategies that actors may use to gain access; and the rules that are followed in the making of publicly binding decisions. To work properly, the ensemble must be institutionalized that is to say, the various patterns must be habitually known, practiced, and accepted by most, if not all, actors. Increasingly, the preferred mechanism of institutionalization is a written body of laws undergirded by a written constitution, though many enduring political norms can have an informal, prudential, or traditional basis. 4 For the sake of economy and comparison, these forms, characteristics, and rules are usually bundled together and given a generic label. Democratic is one; others are autocratic, authoritarian, despotic, dictatorial, tyrannical, totalitarian, absolutist, traditional, monarchic, oligarchic, plutocratic, aristocratic, and sultanistic) Each of these regime forms may in turn be broken down into subtypes. Like all regimes, democracies depend upon the presence of rulers, persons who occupy specialized authority roles and can give legitimate commands to others. What distinguishes democratic rulers from nondemocratic ones are the norms that condition how the former come to power and the practices that hold them accountable for their actions. Schmitter & Karl 77 The public: realm encompasses the making of collective norms and choices that are binding on the society and backed by state coercion. Its content can vary a great deal across democracies, depending upon preexisting distinctions between the public and the private, state and society, legitimate coercion and voluntary exchange, and collective needs and individual preferences. The liberal conception of democracy advocates circumscribing the public realm as narrowly as possible, while the socialist or social-democratic approach would extend that realm through regulation, subsidization, and, in some cases, collective ownership of property. Neither is intrinsically more democratic than the other--just differently democratic. This implies that measures aimed at "developing the private sector" are no more democratic than those aimed at "developing the public sector." Both, if carried to extremes, could undermine the practice of democracy, the former by destroying the basis for satisfying collective needs and exercising legitimate authority; the latter by destroying the basis for satisfying individual preferences and controlling illegitimate government actions. Differences of opinion over the optimal mix of the two provide much of the substantive content of political conflict within established democracies. Citizens are the most distinctive element in democracies. All regimes have rulers and a public realm, but only to the extent that they are democratic do they have citizens. Historically, severe restrictions on citizenship were imposed in most emerging or partial democracies according to criteria of age, gender, class, race, literacy, property ownership, tax-paying status, and so on. Only a small part of the total population was eligible to vote or run for office. Only restricted social categories were allowed to form, join, or support political associations. After protracted struggle--in some cases involving violent domestic upheaval or international war--most of these restrictions were lifted. Today, the criteria for inclusion are fairly standard. All native-born adults are eligible, although somewhat higher age limits may still be imposed upon candidates for certain offices. Unlike the early American and European democracies of the nineteenth century, none of the recent democracies in southern Europe, Latin America, Asia, or Eastern Europe has even attempted to impose formal restrictions on the franchise or eligibility to office. When it comes to informal restrictions on the effective exercise of citizenship rights, however, the story can be quite different. This explains the central importance (discussed below) of procedures. Competition has not always been considered an essential defining condition of democracy. "Classic" democracies presumed decision making based on direct participation leading to consensus. The assembled citizenry was expected to agree on a common course of action after listening to the alternatives and weighing their respective merits and demerits. A tradition of hostility to "faction," and "particular interests" 78 Journal of Democracy persists in democratic thought, but at least since The Federalist Papers it has become widely accepted that competition among factions is a necessary evil in democracies that operate on a more-than-local scale. Since, as James Madison argued, "the latent causes of faction are sown into the nature of man," and the possible remedies for "the mischief of faction" are worse than the disease, the best course is to " H o w e v e r central recognize them and to attempt to control to democracy, their effects.6 Yet while democrats may elections occur agree on the inevitability of factions, they i n t e r m i t t e n t l y and tend to disagree about the best forms and only a l l o w citizens rules for governing factional competition. to choose between Indeed, differences over the preferred the highly modes and boundaries of competition aggregated contribute most to distinguishing one alternatives subtype of democracy from another. offered by political The most popular definition of parties..." democracy equates it with regular elections, fairly conducted and honestly counted. Some even consider the mere fact of elections---even ones from which specific parties or candidates are excluded, or in which substantial portions of the population cannot freely participate--as a sufficient condition for the existence of democracy. This fallacy has been called "electoralism" or "the faith that merely holding elections will channel political action into peaceful contests among elites and accord public legitimacy to the winners" no matter how they are conducted or what else constrains those who win them. 7 However central to democracy, elections occur intermittently and only allow citizens to choose between the highly aggregated alternatives offered by political parties, which can, especially in the early stages of a democratic transition, proliferate in a bewildering variety. During the intervals between elections, citizens can seek to influence public policy through a wide variety of other intermediaries: interest associations, social movements, locality groupings, clientelistic arrangements, and so forth. Modern democracy, in other words, offers a variety of competitive processes and channels for the expression of interests and values---associational as well as partisan, functional as well as territorial, collective as well as individual. All are integral to its practice. Another commonly accepted image of democracy identifies it with majority rule. Any governing body that makes decisions by combining the votes of more than half of those eligible and present is said to be democratic, whether that majority emerges within an electorate, a parliament, a committee, a city council, or a party caucus. For exceptional purposes (e.g., amending the constitution or expelling a member), "qualified majorities" of more than 50 percent may be Schmitter & Karl 79 required, but few would deny that democracy must involve some means of aggregating the equal preferences of individuals. A problem arises, however, when numbers meet intensities. What happens when a properly assembled majority (especially a stable, selfperpetuating one) regularly makes decisions that harm some minority (especially a threatened cultural or ethnic group)? In these circumstances, successful democracies tend to qualify the central principle of majority rule in order to protect minority rights. Such qualifications can take the form of constitutional provisions that place certain matters beyond the reach of majorities (bills of rights); requirements for concurrent majorities in several different constituencies (confederalism); guarantees securing the autonomy of local or regional governments against the demands of the central authority (federalism); grand coalition governments that incorporate all parties (consociationalism); or the negotiation of social pacts between major social groups like business and labor (neocorporatism). The most common and effective way of protecting minorities, however, lies in the everyday operation of interest associations and social movements. These reflect (some would say, amplify) the different intensities of preference that exist in the population and bring them to bear on democratically elected decision makers. Another way of putting this intrinsic tension between numbers and intensities would be to say that "in modern democracies, votes may be counted, but influences alone are weighted." Cooperation has always been a central feature of democracy. Actors must voluntarily make collective decisions binding on the polity as a whole. They must cooperate in order to compete. They must be capable of acting collectively through parties, associations, and movements in order to select candidates, articulate preferences, petition authorities, and influence policies. But democracy's freedoms should also encourage citizens to deliberate among themselves, to discover their common needs, and to resolve their differences without relying on some supreme central authority. Classical democracy emphasized these qualities, and they are by no means extinct, despite repeated efforts by contemporary theorists to stress the analogy with behavior in the economic marketplace and to reduce all of democracy's operations to competitive interest maximization. Alexis de Tocqueville best described the importance of independent groups for democracy in his Democracy in America, a work which remains a major source of inspiration for all those who persist in viewing democracy as something more than a struggle for election and re-election among competing candidates. 8 In contemporary political discourse, this phenomenon of cooperation and deliberation via autonomous group activity goes under the rubric of "civil society." The diverse units of social identity and interest, by remaining independent of the state (and perhaps even of parties), not 80 Journal of Democracy only can restrain the arbitrary actions of rulers, but can also contribute to forming better citizens who are more aware of the preferences of others, more self-confident in their actions, and more civic-minded in their willingness to sacrifice for the common good. At its best, civil society provides an intermediate layer of governance between the individual and the state that is capable of resolving conflicts and controlling the behavior of members without public coercion. Rather than overloading decision makers with increased demands and making the system ungovernable,9 a viable civil society can mitigate conflicts and improve the quality of citizenship---without relying exclusively on the privatism of the marketplace. Representatives--whether directly or indirectly elected-----do most of the real work in modem democracies. Most are professional politicians who orient their careers around the desire to fill key offices. It is doubtful that any democracy could survive without such people. The central question, therefore, is not whether or not there will be a political elite or even a professional political class, but how these representatives are chosen and then held accountable for their actions. As noted above, there are many channels of representation in modem democracy. The electoral one, based on territorial constituencies, is the most visible and public. It culminates in a parliament or a presidency that is periodically accountable to the citizenry as a whole. Yet the sheer growth of government (in large part as a byproduct of popular demand) has increased the number, variety, and power of agencies charged with making public decisions and not subject to elections. Around these agencies there has developed a vast apparatus of specialized representation based largely on functional interests, not territorial constituencies. These interest associations, and not political parties, have become the primary expression of civil society in most stable democracies, supplemented by the more sporadic interventions of social movements. The new and fragile democracies that have sprung up since 1974 must live in "compressed time." They will not resemble the European democracies of the nineteenth and early twentieth centuries, and they cannot expect to acquire the multiple channels of representation in gradual historical progression as did most of their predecessors. A bewildering array of parties, interests, and movements will all simultaneously seek political influence in' them, creating challenges to the polity that did not exist in earlier processes of democratization. Procedures that Make Democracy Possible The defining components of democracy are necessarily abstract, and may give rise to a considerable variety of institutions and subtypes of democracy. For democracy to thrive, however, specific procedural norms Schmitter & Karl 81 must be followed and civic rights must be respected. Any polity that fails to impose such restrictions upon itself, that fails to follow the "rule of law" with regard to its own procedures, should not be considered democratic. These procedures alone do not define democracy, but their presence is indispensable to its persistence. In essence, they are necessary but not sufficient conditions for its existence. Robert Dahl has offered the most generally accepted listing of what he terms the "procedural minimal" conditions that must be present for modem political democracy (or as he puts it, "polyarchy") to exist: 1) Control over government decisions about policy is constitutionally vested in elected officials. 2) Elected officials are chosen in frequent and fairly conducted elections in which coercion is comparatively uncommon. 3) Practically all adults have the fight to vote in the election of officials. 4) Practically all adults have the fight to run for elective offices in the government . . . . 5) Citizens have a fight to express themselves without the danger of severe punishment on political matters broadly defined . . . . 6) Citizens have a fight to seek out alternative sources of information. Moreover, altemative sources of information exist and are protected by law. 7) . . . Citizens also have the fight to form relatively independent associations or organizations, including independent political parties and interest groups. ~~ These seven conditions seem to capture the essence of procedural democracy for many theorists, but we propose to add two others. The first might be thought of as a further refinement of item (1), while the second might be called an implicit prior condition to all seven of the above. 8) Popularly elected officials must be able to exercise their constitutional powers without being subjected to overriding (albeit informal) opposition from unelected officials. Democracy is in jeopardy if military officers, entrenched civil servants, or state managers retain the capacity to act independently of elected civilians or even veto decisions made by the people's representatives. Without this additional caveat, the militarized polities of contemporary Central America, where civilian control over the military does not exist, might be classified by many scholars as democracies, just as they have been (with the exception of Sandinista Nicaragua) by U.S. policy makers. The caveat thus guards against what we earlier called "electoralism"--the tendency to focus on the holding of elections while ignoring other political realities. 9) The polity must be self-governing; it must be able to act independently of constraints imposed by some other overarching political system. Dahl and other contemporary democratic theorists probably took 82 Journal of Democracy this condition for granted since they referred to formally sovereign nation-states. However, with the development of blocs, alliances, spheres of influence, and a variety of "neocolonial" arrangements, the question of autonomy has been a salient one. Is a system really democratic if its elected officials are unable to make binding decisions without the approval of actors outside their territorial domain? This is significant even if the outsiders are themselves democratically constituted and if the insiders are relatively free to alter or even end the encompassing arrangement (as in Puerto Rico), but it becomes especially critical if neither condition obtains (as in the Baltic states). Principles that Make Democracy Feasible Lists of component processes and procedural norms help us to specify what democracy is, but they do not tell us much about how it actually functions. The simplest answer is "by the consent of the people"; the more complex one is "by the contingent consent of politicians acting under conditions of bounded uncertainty." In a democracy, representatives must at least informally agree that those who win greater electoral support or influence over policy will not use their temporary superiority to bar the losers from taking office or exerting influence in the future, and that in exchange for this opportunity to keep competing for power and place, momentary losers will respect the winners' right to make binding decisions. Citizens are expected to obey the decisions ensuing from such a process of competition, provided its outcome remains contingent upon their collective preferences as expressed through fair and regular elections or open and repeated negotiations. The challenge is not so much to find a set of goals that command widespread consensus as to find a set of rules that embody contingent consent. The precise shape of this "democratic bargain," to use Dahl's expression," can vary a good deal from society to society. It depends on social cleavages and such subjective factors as mutual trust, the standard of fairness, and the willingness to compromise. It may even be compatible with a great deal of dissensus on substantive policy issues. All democracies involve a degree of uncertainty about who will be elected and what. policies they will pursue. Even-in those polities where one party persists in winning elections or one policy is consistently implemented, the possibility of change through independent collective action still exists, as in Italy, Japan, and the Scandinavian social democracies. If it does not, the system is not democratic, as in Mexico, Senegal, or Indonesia. But the uncertainty embedded in the core of all democracies is bounded. Not just any actor can get into the competition and raise any issue he or she pleases--there are previously established rules that must 83 Schmitter & Karl be respected. Not just any policy can be adopted there are conditions that must be met, Democracy institutionalizes "normal," limited political uncertainty. These boundaries vary from country to country. Constitutional guarantees of property, privacy, expression, and other rights are a part of this, but the most effective boundaries are generated by competition among interest groups and cooperation within civil society. Whatever the rhetoric (and some polities appear to offer their citizens more dramatic alternatives than others), once the rules of contingent consent have been agreed upon, the actual variation is likely to stay within a predictable and generally accepted range. This emphasis on operative guidelines contrasts with a highly persistent, but misleading theme in recent literature on democracymnamely, the emphasis upon "civic culture." The principles we have suggested here rest on rules of prudence, not on deeply ingrained habits of tolerance, moderation, mutual respect, fair play, readiness to compromise, or trust in public authorities. Waiting for such habits to sink deep and lasting roots implies a very slow process of regime consolidation---one that takes generations--and it would probably condemn most contemporary experiences ex hypothesi to failure. Our assertion is that contingent consent and bounded uncertainty can emerge from the interaction between antagonistic and mutually suspicious actors and that the far more benevolent and ingrained norms of a civic culture are better thought of as a product and not a producer of democracy. H o w Democracies Differ Several concepts have been deliberately excluded from our generic definition of democracy, despite the fact that they have been frequently associated with it in both everyday practice and scholarly work. They are, nevertheless, especially important when it comes to distinguishing subtypes of democracy. Since no single set of actual institutions, practices, or values embodies democracy, polities moving away from authoritarian rule can mix different components to produce different democracies. It is important to recognize that these do not define points along a single continuum of improving performance, but a matrix of potential combinations that are differently democratic. 1) Consensus: All citizens may not agree on the substantive goals of political action or on the role of the state (although if they did, it would certainly make governing democracies much easier). 2) Participation: All citizens may not take an active and equal part in politics, although it must be legally possible for them to do so. 3) Access: Rulers may not weigh equally the preferences of all who come before them, although citizenship implies that individuals and groups should have an equal opportunity to express their preferences if they choose to do so. 84 Journal of Democracy 4) Responsiveness: Rulers may not always follow the course of action preferred by the citizenry. But when they deviate from such a policy, say on grounds of "reason of state" or "overriding national interest," they must ultimately be held accountable for their actions through regular and fair processes. 5) Majority rule: Positions may not be allocated or rules may not be decided solely on the basis of assembling the most votes, although deviations from this principle usually must be explicitly defended and previously approved. 6) Parliamentary sovereignty: The legislature may not be the only body that can make rules or even the one with final authority in deciding which laws are binding, although where executive, judicial, or other public bodies make that ultimate choice, they too must be accountable for their actions. 7) Party government: Rulers may not be nominated, promoted, and disciplined in their activities by well-organized and programmatically coherent political parties, although where they are not, it may prove more difficult to form an effective government. 8) Pluralism: The political process may not be based on a multiplicity of overlapping, voluntaristic, and autonomous private groups. However, where there are monopolies of representation, hierarchies of association, and obligatory memberships, it is likely that the interests involved will be more closely linked to the state and the separation between the public and private spheres of action will be much less distinct. 9) Federalism: The territorial division of authority may not involve multiple levels and local autonomies, least of all ones enshrined in a constitutional document, although some dispersal of power across territorial and/or functional units is characteristic of all democracies. 10) Presidentialism: The chief executive officer may not be a single person and he or she may not be directly elected by the citizenry as a whole, although some concentration of authority is present in all democracies, even if it is exercised collectively and only held indirectly accountable to the electorate. l l) Checks and Balances: It is not necessary that the different branches of government be systematically pitted against one another, although governments by assembly, by executive concentration, by judicial command, or even by dictatorial fiat (as in time of war) must be ultimately accountable to the citizenry as a whole. While each of the above has been named as an essential component of democracy, they should instead be seen either as indicators of this or that type of democracy, or else as useful standards for evaluating the perform~mce of particular regimes. To include them as part of the generic definition of democracy itself would be to mistake the American polity for the universal model of democratic governance. Indeed, the parliamentary, consociational, unitary, corporatist, and concentrated 85 Schmitter & Karl arrangements of continental Europe may have some unique virtues for guiding polities through the uncertain transition from autocratic to democratic rule. ~2 W h a t D e m o c r a c y Is N o t We have attempted to convey the general meaning of modern democracy without identifying it with some particular set of rules and institutions or restricting it to some specific culture or level of development. We have also argued that it cannot be reduced to the regular holding of elections or equated with a particular notion of the role of the state, but we have not said much more about what democracy is not or about what democracy may not be capable of producing. There is an understandable temptation to load too many expectations on this concept and to imagine that by attaining democracy, a society will have resolved all of its political, social, economic, administrative, and cultural problems. Unfortunately, "all good things do not necessarily go together." First, democracies are not necessarily more efficient economically than other forms of government. Their rates of aggregate growth, savings, and investment may be no better than those of nondemocracies. This is especially likely during the transition, when propertied groups and administrative elites may respond to real or imagined threats to the "rights" they enjoyed under authoritarian rule by initiating capital flight, disinvestment, or sabotage. In time, depending upon the type of democracy, benevolent long-term effects upon income distribution, aggregate demand, education, productivity, and creativity may eventually combine to improve economic and social performance, but it is certainly too much to expect that these improvements will occur immediately--much less that they will be defining characteristics of democratization. Second, democracies are not necessarily more efficient administratively. Their capacity to make decisions may even be slower than that of the regimes they replace, if only because more actors must be consulted. The costs of getting things done may be higher, if only because "payoffs" have to be made to a wider and more resourceful set of clients (although one should never underestimate the degree of corruption to be found within autocracies). Popular satisfaction with the new democratic government's performance may not even seem greater, if only because necessary compromises often please no one completely, and because the losers are free to complain. Third, democracies are not likely to appear more orderly, consensual, stable, or governable than the autocracies they replace. This is partly a byproduct of democratic freedom of expression, but it is also a reflection of the likelihood of continuing disagreement over new rules and 86 Journal of Democracy institutions. These products of imposition or compromise are often initially quite ambiguous in nature and uncertain in effect until actors have learned how to use them. What is more, they come in the aftermath of serious struggles motivated by high ideals. Groups and individuals with recently acquired autonomy will test certain rules, protest against the actions of certain institutions, and insist on "'...democracies renegotiating their part of the bargain. Thus will have the presence of antisystem parties should more open societies and be neither surprising nor seen as a failure of democratic consolidation. What counts is polities than the whether such parties are willing, however autocracies reluctantly, to play by the general rules of they replace, but bounded uncertainty and contingent consent. n o t necessarily Governability is a challenge for all more open regimes, not just democratic ones. Given economies." the political exhaustion and loss of legitimacy that have befallen autocracies from sultanistic Paraguay to totalitarian Albania, it may seem that only democracies can now be expected to govern effectively and legitimately. Experience has shown, however, that democracies too can lose the ability to govern. Mass publics can become disenchanted with their performance. Even more threatening is the temptation for leaders to fiddle with procedures and ultimately undermine the principles of contingent consent and bounded uncertainty. Perhaps the most critical moment comes once the politicians begin to settle into the more predictable roles and relations of a consolidated democracy. Many will find their expectations frustrated; some will discover that the new rules of competition put them at a disadvantage; a few may even feel that their vital interests are threatened by popular majorities. Finally, democracies will have more open societies and polities than the autocracies they replace, but not necessarily more open economies. Many of today's most successful and well-established democracies have historically resorted to protectionism and closed borders, and have relied extensively upon public institutions to promote economic development. While the long-term compatibility between democracy and capitalism does not seem to be in doubt, despite their continuous tension, it is not clear whether the promotion of such liberal economic goals as the right of individuals to own property and retain profits, the cleating function of markets, the private settlement of disputes, the freedom to produce without government regulation, or the privatization of state-owned enterprises necessarily furthers the consolidation of democracy. After all, democracies do need to levy taxes and regulate certain transactions, especially where private monopolies and oligopolies exist. Citizens or their representatives may decide that it is desirable to protect the rights 87 Schmitter & Karl of collectivities from encroachment by individuals, especially propertied ones, and they may choose to set aside certain forms of property for public or cooperative ownership. In short, notions of economic liberty that are currently put forward in neoliberal economic models are not synonymous with political f r e e d o m - - a n d may even impede it. Democratization will not necessarily bring in its wake economic growth, social peace, administrative efficiency, political harmony, free markets, or "the end of ideology." Least of all will it bring about "the end o f history." No doubt some of these qualities could make the consolidation o f democracy easier, but they are neither prerequisites for it nor immediate products of it. Instead, what we should be hoping for is the emergence of political institutions that can peacefully compete to form governments and influence public policy, that can channel social and economic conflicts through regular procedures, and that have sufficient linkages to civil society to represent their constituencies and commit them to collective courses of action. Some types of democracies, especially in developing countries, have been unable to fulfill this promise, perhaps due to the circumstances o f their transition from authoritarian rule. 13 The democratic wager is that such a regime, once established, will not only persist by reproducing itself within its initial confining conditions, but will eventually expand beyond them. 14 Unlike authoritarian regimes, democracies have the capacity to modify their rules and institutions consensually in response to changing circumstances. They may not immediately produce all the goods mentioned above, but they stand a better chance o f eventually doing so than do autocracies. NOTES 1. For a comparative analysis of the recent regime changes in southern Europe and Latin America, see Guillermo O'Donnell, Philippe C. Schmitter, and Laurence Whitehead, eds., Transitions from Authoritarian Rule, 4 vols. (Baltimore: Johns Hopkins University Press, 1986). For another compilation that adopts a more structural approach see Larry Diamond, Juan Linz, and Seymour Martin Lipset, eds., Democracy in Developing Countries, vols. 2, 3, and 4 (Boulder, Colo.: Lynne Rienner, 1989). 2. Numerous attempts have been made to codify and quantify the existence of democracy across political systems. The best known is probably Freedom House's Freedom in the WorM: Political Rights and Civil Liberties, published since 1973 by Greenwood Press and since 1988 by University Press of America. Also see Charles Humana, Worm Human Rights Guide (New York: Facts on File, 1986). 3. The definition most commonly used by American social scientists is that of Joseph Schumpeter: "that institutional arrangement for arriving at political decisions in which individuals acquire the power to decide by means of a competitive struggle for the people's vote." Capitalism, Socialism and Democracy (London: George Allen and Unwin, 1943), 269. We accept certain aspects of the classical procedural approach to modem democracy, but differ primarily in our emphasis on the accountability of rulers to citizens and the relevance of mechanisms of competition other than elections. 4. Not only do some countries practice a stable form of democracy without a formal constitution (e.g., Great Britain and Israel), but even more countries have constitutions and 88 Journal of Democracy legal codes that offer no guarantee of reliable practice. On paper, Stalin's 1936 constitution for the USSR was a virtual model of democratic rights and entitlements. 5. For the most valiant attempt to make some sense out of this thicket of distinctions, see Juan Linz, "Totalitarian and Authoritarian Regimes" in Handbook of Political Science, eds. Fred I. Greenstein and Nelson W. Polsby (Reading, Mass.: Addision Wesley, 1975), 175-411. 6. "Publius" (Alexander Hamilton, John Jay, and James Madison), The Federalist Papers (New York: Anchor Books, 1961). The quote is from Number 10. 7. See Terry Karl, "Imposing Consent? Electoralism versus Democratization in E1 Salvador," in Elections and Democratization in Latin America, 1980-1985, eds. Paul Drake and Eduardo Silva (San Diego: Center for Iberian and Latin American Studies, Center for US/Mexican Studies, University of California, San Diego, 1986), 9-36. 8. Alexis de Tocqueville, Democracy in America, 2 vols. (New York: Vintage Books, 1945). 9. This fear of overloaded government and the imminent collapse of democracy is well reflected in the work of Samuel P. Huntington during the 1970s. See especially Michel Crozier, Samuel P. Huntington, and Joji Watanuki, The Crisis of Democracy (New York: New York University Press, 1975). For Huntington's (revised) thoughts about the prospects for democracy, see his "Will More Countries Become Democratic?," Political Science" Quarterly 99 (Summer 1984): 193-218. 10. Robert Dahl, Dilemmas of Pluralist Democracy (New Haven: Yale University Press, 1982), 11. 11. Robert Dahl, After the Revolution: Authority in a Good Society (New Haven: Yale University Press, 1970). 12. See Juan Linz, "The Perils of Presidentialism," Journal of Democracy 1 (Winter 1990): 51-69, and the ensuing discussion by Donald Horowitz, Seymour Martin Lipset, and Juan Linz in Journal of Democracy 1 (Fall 1990): 73-91. 13. Terry Lynn Karl, "Dilemmas of Democratization in Latin America," Comparative Politics 23 (October 1990): 1-23. 14. Otto Kirchheimer, "Conf'ming Conditions and Revolutionary Breakthroughs," American Political Science Review 59 (1965): 964-974.