Framing the Research Questions

One element of this new baseline concerns the outcome to be explained. In this regard, this new literature helped to cement an emerging consensus regarding the concept of democracy. The question “What is democracy?” is a matter of continuing discussion, as I will address below. Nonetheless, building on Dahl's earlier work, O'Donnell and Schmitter's emphasis on the “consensus in the contemporary world” regarding the status of a “procedural minimum” as a “necessary” feature of political democracy was valuable.Footnote ⁶ It helped to specify the outcome that called for an explanation in terms that made sense to researchers interested in a diverse set of countries at a time when doubts about such a procedural minimum were still raised in some quarters. And it provided a much needed conceptual anchor for empirical research.

Relatedly, this new literature helped establish the importance of disaggregating the problematic of democracy into distinct questions in an organized fashion. At the broadest level, understanding the conditions of democracy was seen as hinging on two distinct questions: (1) What are the conditions for a transition from some form of authoritarianism to democracy? and (2) What are the factors that account for the durability or endurance of democracy? In turn, the first question was disaggregated into two questions, about the conditions for a transition from authoritarianism and for a transition to democracy. That is, the overarching research question—What explains democracy?—was broken down into component parts that were both amenable to more careful analysis and explicitly linked together.

Again, these were not new ideas. In particular, Dankwart Rustow had earlier introduced the distinction between a theory of democratic transitions and a theory of the stability of democracy.Footnote ⁷ But, in stressing the possible asymmetry between processes of transition to democracy and the breakdown of democracy, O'Donnell and Schmitter highlighted the need to consider the possibility that questions about the origins and the durability of democracy might call for different answers. Moreover, by underscoring that a theory of the demise of authoritarian rule is not, by itself, a theory of democratic transitions, and by clarifying that liberalization is not the same as democratization and can result rather in a liberalized autocracy (dictablanda) or a limited democracy (democradura), O'Donnell and Schmitter spelled out the explanatory challenges with uncanny lucidity.Footnote ⁸

While O'Donnell and Schmitter's way of framing the research on democratization has not, to be sure, been adopted by all, it dramatically transformed scholarship on democracy. The consensus around the procedural minimum of democracy has broadened. In addition, more and more scholars distinguish between processes of transition to democracy and breakdowns of democracy. For example, Ruth Berins Collier explicitly distinguishes the origins from the endurance of democracy and has devoted considerable attention to the task of identifying the threshold countries must cross in a democratic transition.Footnote ⁹ Charles Tilly similarly places the distinction between processes of democratization and “de-democratization” at the heart of his analysis, and repeatedly highlights the asymmetry between these two processes.Footnote ¹⁰ And Daron Acemoglu and James Robinson's “unified theory of democracy” centers on the challenge of providing an explanation that can account for both the origins and the durability of democracy.Footnote ¹¹ At the same time, the importance of distinguishing transitions from authoritarianism and transitions to democracy is at the heart of the large new literature on “hybrid regimes,” that is, regimes that combine features of authoritarianism and democracy and thus occupy a grey zone between clearly authoritarian and democratic systems.Footnote ¹² In all of these respects, the impact of the Woodrow Wilson Center project on transitions on the framing of research was profound.

Theorizing Political Dynamics

A second element of the new baseline established in the new phase of research on democracy concerns the development of explanatory theories. The literature on democratic transitions, and in particular O'Donnell and Schmitter's book and Przeworski's early work on transitions,Footnote ¹³ broke with the focus on macrosocietal forces that figured prominently in the earlier works of Lipset and Moore and in other classic books.Footnote ¹⁴ The reason for this shift in approach, as the instigators of this shift have acknowledged,Footnote ¹⁵ was largely political rather than purely scientific. In Przeworski's words, in the older literature, outcomes were “uniquely determined by [macrosocietal] conditions, and history … [went] on without anyone ever doing anything.”Footnote ¹⁶ That is, the prevailing macrosocietal theories of the 1960s and 1970s were seen as overly deterministic and hence offering no guidance to political actors or, even worse, seemed to invite quiescence. Indeed, the political reading of these theories offered by some scholars highlighted that many authoritarian countries lacked the supposed prerequisites for democracy and that political activity oriented toward promoting democracy when a country was not ripe for democracy might actually be counterproductive. The participants in the Wilson Center project on transitions thus practiced what Abraham Lowenthal characterized as “thoughtful wishing,”Footnote ¹⁷ rejecting “structural determinism,” unabashedly embracing democracy as a positive normative value, and searching for realistic possibilities to foster democracy.

If the break with prevailing macrosocietal theories in this new literature was largely politically driven, the alternative theoretical approach that these writers offered stands on its own merits. The new theorizing focused on process, drew attention to both state actors and societal actors, and analyzed the choices made by these actors. Rejecting structuralist determinism, O'Donnell and Schmitter argued that political outcomes were “underdetermined” from the perspective of standard macrostructural factors, since political outcomes were contingent on the strategic choices of actors and, most critically, because key characteristics of the relevant actors such as their power were not givens but rather were affected by their choices and underwent significant changes in the course of the political process itself.Footnote ¹⁸

In developing these ideas, O'Donnell and Schmitter drew on Rustow's earlier work on transitions to democracy and Linz's analysis of the breakdown of democracy.Footnote ¹⁹ Przeworski, too, had earlier stressed, in his analysis of social democracy, that “Classes are not a datum prior to the history of concrete struggles” and that “politics … should be seen as a contingent historical outcome of continual conflicts.”Footnote ²⁰ But the Wilson Center project brought these insights into greater prominence. The term “endogeneity,” used to denote explanatory factors that are affected by the variables they are purported to affect, did not enter the vocabulary of students of democracy until later, in the mid-1990s. But it is not a stretch to say that the new literature on democratic transitions was grappling with the problem of endogeneity and suggesting, as Przeworski put it later, that “everything is endogenous” and hence that “[t]he only motor of history is endogeneity.”Footnote ²¹

The influence of this turn to political explanations is most obvious in recent institutionalist scholarship that treats political institutions as results of the decisions of political actors.Footnote ²² It is also evident in recent research that revisits some of the classic macrosocietal arguments of the earlier literature on democracy. For example, when Collier revisited the debate about classes and democracy launched by Moore, she explicitly included political elites, both within and outside the state, in her explanatory framework.Footnote ²³ Similarly, attempts to revisit the classic macroeconomic and macrosocial explanations of the 1960s using the tools of game theory highlight the strategic interaction among actors and the choices that elites in particular make in light of the redistributive consequences of state policies.Footnote ²⁴ These developments mark the emergence of a new political sociology and political economy of democracy, more attuned to the duality of structure and agency and to the endogeneity of societal forces and actors.Footnote ²⁵

Testing Propositions

Finally, a third element of the new baseline established in the new research on democracy concerns the empirical testing of propositions and, relatedly, the use of the results of empirical research for the purpose of offering policy advice. The literature of the 1960s and 1970s was empirical. But few authors used cross-national data, provided any sort of rigorous statistical test of their explanatory propositions, or paid much attention to matters of research design when drawing conclusions about the causes of democracy. These limitations, however, did not stop authors such as Samuel Huntington from mobilizing arguments, among other things about the economic prerequisites of democracy, to support authoritarianism during the Cold War.Footnote ²⁶

These characteristics of the literature on democracy began to change, in large part due to the work of Przeworski, starting with his 1997 World Politics article on modernization theory and his 2000 book Democracy and Development. Footnote ²⁷ These publications showed that by investing efforts in data generation, being attentive to endogeneity, and drawing on relevant statistical techniques, it was possible to assess arguments in a much more rigorous manner than was the case in earlier decades. They also showed how careful academic research could be used to unmask influential ideological arguments about democracy and offer a more responsible basis for debate about policy choices.

Specifically, Przeworski tackled the argument that, because economic development supposedly leads to democracy and authoritarian governments ostensibly perform better in economic terms, the wise thing to do is support authoritarian governments and their development projects. He showed, based on a more scientifically valid evaluation than had ever previously been conducted, that “we do not have a shred of evidence that democracy need be sacrificed on the altar of development” and that “democracy is … neither inevitable nor impossible … [but rather] a contingent outcome of conflicts among several organized groups.”Footnote ²⁸ Przeworski thus added force to the rejection of the thesis about the economic prerequisites of democratization offered by the Wilson Center project, mainly based on political considerations, by questioning the empirical validity of the sort of political arguments Huntington was known for.Footnote ²⁹

Much as with other changes introduced or consolidated during this phase of research on democracy, not all writers on democracy have embraced the new standard concerning the testing of propositions, especially when seeking to extract policy recommendations from research. Perhaps more worrisome, several authors, including public intellectuals Fareed Zakaria and Francis Fukuyama, have recently resurrected Huntingtonian arguments cautioning against the “premature” promotion of democratization.Footnote ³⁰ While such calls for postponing, demoting or containing demands for democracy have increasingly placed emphasis on the need to secure political order and not just economic development before pushing for democracy,Footnote ³¹ these claims about prerequisites of democracy are typically based on a clearly selective reading of the evidence and certainly nothing that approximates a robust causal finding.

But such impressionistic arguments fly in the face of an increasingly sophisticated body of research on democratization driven by the commitment to collect cross-national data and subject explanatory claims to a systematic empirical analysis.Footnote ³² A heightened attention to questions of measurement and data generation has been at the root of many important advances in proposition testing. Equally crucial is the increased emphasis on matters of research design, which Robinson refers to as the “causality revolution,” which has gradually spread from the discipline of economics to the study of democracy.Footnote ³³ And though this more methodologically self-conscious literature has not addressed prescriptive questions and political arguments as much as it could and should, it implicitly exposes the purely ideological nature of much advice offered in the name of promoting democracy.

State and Nation in Explanations of Democracy

The relationship between the character of the state and democratization has long been central to the collaboration of Linz and Stepan. These two scholars did some of the pioneering work on the topic in the 1990s, when the fate of two ethnofederal states, the USSR and Yugoslavia, made it abundantly clear that the state could not be treated as a constant and that it was necessary to theorize the relationship between the state, the nation, and democracy.Footnote ³⁶ These authors have dedicated sustained attention to this complex relationship over the past fifteen years, gradually broadening the empirical scope of their analysis and increasingly highlighting the role of federal institutions.Footnote ³⁷ The result of these efforts is their new book on the topic, coauthored with Yogendra Yadav of the Centre of Developing Societies in Delhi: Crafting State-Nations: India and Other Multinational Democracies.

Linz and Stepan argue that when a process of democratization starts to unfold in a society that is multinational, political demands will likely not focus solely on altering the way key government offices are accessed. Questions will also be raised about the appropriateness of the political unit, that is, over what territory and what population a state should govern. Thus, discussions about democracy and the state cannot be divorced. Yet the standard answer—that only nation-states, that is, states that contain only one nation, can generate the necessary degree of trust in the major institutions of the state that a modern democracy needs, and hence every state should be a nation and every nation should be a state—is problematic. When a society is multinational, building a nation-state is likely to be incompatible with developing democracy, requiring the marginalization of important sectors of the population and even a heavy dose of repression. Moreover, the difficulty of creating nation-states through state policies has been increasingly demonstrated in the twentieth century.

Seeking a way out of this quandary, Linz and Stepan offer some interesting ideas. First, they introduce the concept of a “state-nation,” characterized by multiple but complementary political identities and loyalties, a high degree of positive identification with the state, and pride in being citizens of that state, which they see as more appropriate than the idea of the nation-state for culturally diverse societies and multinational societies in particular. They also suggest that democratization need not be incompatible with a questioning of the appropriateness of the political unit if multinational societies opt for asymmetrical federalism, an arrangement in which some cultural prerogatives are constitutionally embedded for subunits with salient and mobilized territorial identities, along with other consociational features.

In short, emphasizing institutional choices in a manner that is reminiscent of Lijphart,Footnote ³⁸ Linz and Stepan develop a theoretical argument that counters Rustow's view that national unity is a key prerequisite of democracy.Footnote ³⁹ The Rustow thesis, Linz and Stepan's analysis implies, is based on an unduly restrictive model of the state that does not envision the real possibility that questions regarding the political unit need not be settled before democratization but, instead, can be tackled simultaneously. That is, Linz and Stepan show how the question of the state can be endogenized within a theory of democratization rather than treated as a fixed background condition.

Linz and Stepan's ideas about the relationship between the state and democracy do not settle the question of how to incorporate the state within a theory of democracy, let alone arguments about whether the state is a prerequisite of democracy. The viability of their alternative depends to a large extent on the independent power of institutions, a matter that they do not directly address.Footnote ⁴⁰ Moreover, when Linz and Stepan focus on dimensions of the broad problematic of the state that go beyond the appropriateness of the political unit, they more or less explicitly posit that the state still is, in some senses, a prerequisite of democracy. However, these other arguments, introduced without much elaboration, are either unclear or questionable.

For example, Linz's statement that “without a state there can be no democracy” seems to be an overreach.Footnote ⁴¹ Indeed, as Schmitter has noted, “it is an exaggeration to claim that without a state there can be no democracy.”Footnote ⁴² And Linz's statement is also problematic from a theory-building perspective. We might interpret it as referring to a definitional matter, meaning that inasmuch as democracy refers to, at the very least, the access to government offices, democracy presupposes, as a matter of conceptual logic, a state.Footnote ⁴³ But we might also interpret it as a claim about the causes or consequences of democracy. On a critical matter of theory, we are left wondering.

It is also noteworthy that Linz and Stepan explicitly argue that the state is a prerequisite of democracy in the sense that “democracy requires statehood,” that is, the recognition of a territorial unit as a sovereign state.Footnote ⁴⁴ But this claim is challenged, on empirical grounds, by the growing literature on the international dimensions of democratization. As Schmitter remarks, this literature, and especially the strand that studies the vast array of activities that have as a stated aim the international promotion of democracy, is not highly theoretical—another reason to be dubious about the strong prescriptions offered by some authors regarding what Western governments should do and not do in the name of democracy.Footnote ⁴⁵ Nonetheless, it shows that starting in the 1990s, in cases such as Cambodia, Bosnia, Eastern Slavonia, Kosovo, and East Timor, the international community—and, more specifically, the UN—assumed some or all of the governmental powers of a territory and treated the transfer of power from an international administration to an elected government as the key mechanism of an exit strategy.

To be sure, there is an inescapable contradiction involved in efforts to impose democracy from the outside through such international administrations of a territory. After all, the inhabitants of these territories under international administration did not elect their administrators. But the essential point is that, in some cases, the restoration or the gaining of sovereignty is the result of a process, not its point of departure. As the experience in Kosovo shows, not only can democratization proceed when the political status of a territory is a central unresolved issue. In addition, democratization can be fueled precisely by the aim of establishing a claim for sovereignty. Thus, there is such a thing as democratization without a state.Footnote ⁴⁶

To sum up, Linz and Stepan shed important light on the challenges of constructing democracy in the context of multinational and, more broadly, multicultural societies. But there is still much work to do done on this topic. Much hinges on how the relationship between the concepts of the state and democracy is understood. Yet scholars have barely begun to grapple with the various ways in which the state—or more accurately, dimensions of the state—might be considered as causes of democracy, as consequences of democracy, and/or, yet another option (raised next), as constitutive of democracy. Moreover, work on data collection and analysis is sorely needed to turn this discussion into more than a debate about theory. In this regard, the paucity of empirical analyses of the state compared to analyses of modernization theory is striking. While the role of the state is thus increasingly addressed in research on democracy, more theoretical and empirical research is needed to develop this relatively new line of inquiry.

Conceptualizing and Valuing Democracy

The second new line of inquiry that extends the agenda of the Wilson Center project involves a reconsideration of the concept of democracy and its normative value. In the early 1990s, O'Donnell had already suggested that because the conventional view on democracy is based on the experience of the rich democratic countries of Western Europe and North America, the spread of democracy well beyond this region in the last quarter of the twentieth century gave urgency to a reconsideration of standard conceptualizations of democracy. Since then, he has tackled that challenge in a series of essays in which he stressed, among other things, how the state and its multiple dimensions is presupposed by democracy.Footnote ⁴⁷ This long process of reflection has now culminated in his book Democracy, Agency, and the State.

Along similar lines, Przeworski has responded to the perception that the group of scholars that participated in the Wilson Center project “studied transitions to democracy without asking questions about democracy” (p. xii) by turning his attention to conceptualizations of democracy. Indeed, since the 1990s, he has considered the meaning of democracy but also the distinct question of what should be expected from democracy and what, if any, are the limits of democracy.Footnote ⁴⁸ And, much as in O'Donnell's case, this sustained research over many years has been rounded off with a book, Democracy and the Limits of Self-Government.

O'Donnell's and Przeworski's books are alike in many respects. They both rekindle the link between democracy studies and political and legal theory. They both ground their analyses in sweeping historical overviews of the development of democracy. And they both are so substantial that they invite comparison with Dahl's earlier landmark work on the meaning and value of democracy. But these two books also differ in ways that have significant implications for research and for policy debates about democracy. Indeed, it is instructive to compare, if in a necessarily superficial manner, the responses O'Donnell and Przeworski offer to two fundamental questions: What is democracy? And why should democracy be valued? And, because the work of these two authors invites a comparison with Dahl, it is also informative to draw attention to their similarities and differences with Dahl.

Regarding the concept of democracy, the most obvious difference is that between O'Donnell and Dahl. O'Donnell deliberately seeks to broaden the meaning of democracy as elaborated by Dahl.Footnote ⁴⁹ He reframes Dahl's operational, pared-down concept of polyarchy as corresponding to what he labels the democratic regime, and then argues that a full understanding of democracy must reach beyond the democratic regime and encompass, on the one hand, a distinct set of rights and freedoms related to the civil, social and cultural aspects of citizenship and, on the other hand, the legal backing of the agency of all citizens. But O'Donnell's view does not necessarily clash with Dahl's. After all, Dahl not only goes beyond a strictly procedural conception by introducing the idea that certain rights traditionally understood as civil rights are integral to democracy.Footnote ⁵⁰ In addition, he raises the possibility that there are certain features—such as vastly unequal economic resources—that are external to the democratic process yet which can affect the proper functioning of democracy.Footnote ⁵¹ And, going even further, Dahl holds that it is not out of the question to consider whether the very principle that justifies democracy—the principle of equal intrinsic worth in Dahl's case, and the principle of agency in the case of O'Donnell (pp. 25–49)—is upheld in realms that are not directly linked to the democracy process.Footnote ⁵² Dahl's and O'Donnell's approaches thus offer an interesting dialogue on ways of extending the conception of democracy.

Przeworski's concept of democracy is closer to Dahl's than O'Donnell's. His four key conditions of democracy—equality, participation, representation, and liberty—resemble Dahl's quite closely. Moreover, in contrast to O'Donnell, Przeworski keeps the focus squarely on the process of collective decision making and on political institutions rather than rights. But Przeworski moves well beyond a Schumpeterian minimalist concept of democracy and even Dahl's concept of polyarchy. Rather than focus only on whether access to government offices occurs through competitive elections, Przeworski considers what governments can do and cannot do, and stresses how supermajority rule protects vested interests. He places a heavy emphasis on effective equal opportunities and the conditions of political equality. And his warning that “we may be seeing a new monster: democracy without effective citizenship” (p. xiv) resonates with much of O'Donnell's writings. In short, though the differences between O'Donnell and Przeworski are significant, they both propose concepts of democracy that go beyond the essentially minimalist concept used in the Wilson Center project on transitions.

Regarding the value of democracy, Dahl and O'Donnell hold roughly similar views. Dahl argues that a comprehensive set of benefits accrue to democracies. And O'Donnell, as might be expected in light of his focus on agency and emphasis on rights, sees democracy as having a number of benefits beyond the strictly political arena. Nonetheless, Przeworski's position is the most telling, in that he makes an assessment of the value of democracy a central research goal and considers various possible benefits of democracy both from a theoretical and empirical perspective.Footnote ⁵³ And, interestingly, Przeworski's conclusion about the potential virtues of democracy differs considerably from Dahl's and O'Donnell's. Democracy, according to Przeworski, is beneficial because it offers a peaceful way of processing conflicts and embodies the ideal of self-government of the people. But, he underscores, choosing rulers by elections does not lead to greater economic growth and does not assure rationality, representation, or social and economic equality. That is, Przeworski advances a considerably more modest rationale for supporting democracy than either Dahl or O'Donnell.

Przeworski also offers a political counterpoint to O'Donnell. O'Donnell (pp. 209–10) stresses that, though democracy is preferable to its alternatives simply because it provides a peaceful way of deciding who will govern, the normative appeal of democracy extends beyond this basic value. In contrast, Przeworski counters the common tendency to think naively about the promises of democracy by drawing attention to the limits of democracy, cautions against holding democracy up to a standard no other system meets and criticizing democracy for what it cannot realistically deliver, and calls for a discussion of and action on feasible reforms. In other words, a difference between these two scholars—which actually makes their analyses complementary rather than contradictory—is that O'Donnell is especially attentive to the way political ideals serve as a motivation for political action and Przeworski is particularly concerned with offering a guide for political action shorn of unreasonable expectations.

O'Donnell and Przeworski's calls for a broadening of the agenda that has typically occupied students of democracy have major implications. It is worth recalling that most of the research that seeks to explain democracy relies on a minimalist concept of democracy centered on the formation of government. This approach ignores matters such as the ability of a majority to change the status quo and the role of money in politics, is fully divorced from substantive as opposed to procedural considerations, and ignores the ways in which the state itself might be considered constitutive of democracy. Yet these issues—frequently referred to as matters pertaining to the “quality of democracy”—are increasingly moving to the center of the political agenda in many countries, raising new questions about political strategy and complex tradeoffs among multiple values. Thus, this new line of inquiry invites students of democracy to tackle two big challenges: the development of a new consensus about the conceptual and normative underpinning of research on democracy, and the generation of ideas about the new empirical and prescriptive questions that are inevitably raised by the adoption of a concept of democracy that goes beyond the minimalist concept used in the Wilson Center project and indeed in most of the empirical literature on democracy.