P 99 - 004
DEMOCRATIC THEORY AND
COMPARATIVE POLITICS
GUILLERMO O’DONNELL
To my daughter Julia, for the metonymy - and much love
Wissenschaftszentrum Berlin für Sozialforschung gGmbH (WZB)
Reichpietschufer 50, D-10785 Berlin
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Guidelines and tips
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community
Ask the community for help and clear up your study doubts
University Rankings
Discover the best universities in your country according to Docsity users
Free resources
Our save-the-student-ebooks!
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
Democratic Theory and Comparative Politics, Study notes of Theories of Democracy
There are discribes on componenets of a democratic regime, or polyarchy, or poltical democracy and political freedoms from the university of notre dam.
Typology: Study notes
2021/2022
1 / 72
This page cannot be seen from the preview
Don't miss anything!
Related documents
Partial preview of the text
Download Democratic Theory and Comparative Politics and more Study notes Theories of Democracy in PDF only on Docsity!
P 99 - 004
D EMOCRATIC T HEORY AND
C OMPARATIVE P OLITICS
G UILLERMO O ’ D ONNELL
To my daughter Julia, for the metonymy - and much love
Wissenschaftszentrum Berlin für Sozialforschung gGmbH (WZB) Reichpietschufer 50, D-10785 Berlin
Abstract
The present text is, basically, a revision of democratic theory from the perspective of its inadequacies for including into its scope many of the recently democratized countries, as well as some older democracies located outside of the Northwestern quadrant of the world. After warning that it is a first step in a larger and more ambitious endeavor, the paper begins by critically examining various definitions of democracy, especially those that, claiming to follow Schumpeter, are deemed to be "minimalist", or "procesualist." On this basis, a realistic and restricted, but not minimalist, definition of a democratic regime is proposed. After this step, the connections of this topic with several others are explored, including political, social, and welfare rights; the state, especially in its legal dimension; and some characteristics of the overall social context. The main grounding factor that results from these explorations is the conception of agency, especially as it is expressed in the legal system of existing democracies - although the effectiveness of this system and of its underlying conceptions of agency vary quite widely across cases. The approach of the text emphasizes legal and historical factors, while also tracing, in several comparative excursi, some important differences among various kinds of cases. The main conclusions are stated in several propositions, the major thrust of which entails an invitation toward a theoretically disciplined broadening of the analytical and comparative scope of contemporary democratic theory.
Table of Contents
- A Personal Note
- Introduction
- Schumpeter's footnote
- Elections under a Democratic Regime
- Comparative excursus (1)
- or Polyarchy, or Political Democracy 6. On the Components of a Democratic regime,
- Comparative excursus (1)
- A First Look at Political Freedoms
- An institutionalized wager
- Agency and Rights
- The Legal, Pre-Political Construction of Agency
- Comparative excursus (2)
- Political Citizenship and its Correlates
- Comparative excursus (3)
- "Political" freedoms?
- On the state and its legal dimension
- A Look at the Overall Social Context
- Comparative excursus (4)
- Some final propositions
- A Final Pointer
- Coda
- References
D emocratic T heory and C omparative P olitics
Like these streams of the literature, my early work on new democracies 3 was based on the two already mentioned assumptions: there is a clear and consistent corpus of democratic theory, and it can travel well. The problem - my problem, at least - is that I became convinced that the first assumption is wrong and that, by implication, the second one is impracticable. This was a disturbing conclusion; it deprived me of the lenses with which I thought I could immediately put myself to the study of new democracies. Rather, I found myself undertaking a long intellectual detour, during which I internalized, so to speak, my readings about democracy; also, for reasons that will be clear below, I went back to old interests of mine in philosophical, moral, and legal theory.
Another part of this detour was to undertake, under the institutional umbrella of the Kellogg Institute for International Studies of the University of Notre Dame, a series of collaborative studies. These studies were geared to topics that I found important for clarifying empirical and theoretical peculiarities of new, and not so new, democracies, especially but not exclusively in Latin America. One of these projects took stock of the overall situation of democracy in the early 1990s, in both North and South America.^4 Another project looked at Latin America in terms of its pervasive poverty and deep inequality.^5 Still another project studied various aspects of the workings of the legal systems in this region; about the conclusions of this study, suffice to mention that we changed the title of the resulting volume, 6 from "The Rule of Law in Latin America…" to "The (Un)Rule of Law…".
This detour led me to some conclusions, which it may be useful to summarize:
A. An adequate theory of democracy must specify the historical conditions of emergence of various groups of cases or, equivalently, a theory of democracy
(^3) O’Donnell 1992. This text, first published in Brazil in 1988, shares many of the views about
4 ”democratic consolidation”^ that later on I concluded are wrong. I coordinated this project jointly with Abraham Lowenthal. Its main product are papers published in a special series of the Kellogg Institute, where they are available upon request (Castañeda, 5 Conaghan, Dahl, Karl, and Mainwaring, all 1994) This project, coordinated by Víctor Tokman and myself, resulted in the edited book Tokman and 6 O’Donnell 1998. Méndez, O’Donnell, and Pinheiro 1999. This project was coordinated by the coeditors of this volume. Other projects of the Kellogg Institute with which I was less directly involved but from which I greatly benefited were: one that studied the situation after democratization of the welfare systems and social policies in Argentina, Brazil, Chile and Uruguay; and another that studied the present situation of children, especially poor ones, in Latin America. These projects will also resulted in books, presently in the process of publication; see, respectively, Ippolito-O’Donnell, and Bartell and A. O’Donnell, both forthcoming.
G uillermo O ’ D onnell
must include a historically-oriented political sociology of democracy.^7
B. A theory about any social object should not omit within its scope the linguistic uses of its object. The term democracy since times immemorial has been assigned strong, albeit different, moral connotations, all of them grounded in a view of citizens as agents. This opens up democratic theory, including one that is empirically oriented, to complicated but unavoidable issues of political philosophy and moral theory.
C. A theory of democracy, of democracy tout court , must also, and very centrally, include various aspects of legal theory, insofar as the legal system enacts and backs fundamental aspects of both agency and democracy.
D. The preceding points imply that democracy should be analyzed not only at the level of the regime. In addition, it must be studied in relation to the state - especially of the state qua legal system - and to certain aspects of the overall social context.
These conclusions inform texts I wrote during the past ten years, focused on examining characteristics of some new democracies. 8 These characteristics can hardly be deemed as transitory or as just marginally different from what would be expected from existing theories of democracy. In these texts I criticize studies that uncritically "export" such theories. 9 In writings, however, I tackle few themes at a time and then go back to broader issues of democratic theory, but without attempting to discuss or reconstruct this theory as such. Now I feel that I must do this, by means of a book presently in progress. The present text is a preliminary rendering of the first two chapters of this book. It is about democratic theory tout court , aimed at the indispensable goal of clearing conceptual ground for future, more ambitious incursions. However, the intellectual origins of the present text in the study of new democracies will be apparent in some comparative excursus I shall undertake.
(^7) Among works in this perspective that fully or partially focus on Latin America, stand out the excellent contributions of Collier and Collier (1991) and Rueschemeyer, Huber Stephens, and Stephens (1992). But much remains to be done, both in relation to Latin America and its comparison 8 with other regions of the world. 9 O’Donnell 1993, 1994, 1996a, 1996b; the first three of these texts are collected in O’Donnell 1999a. Sartori 1995 has also criticized this procedure; however, our views about how to tackle the resulting problems differ.
G uillermo O ’ D onnell
existing definitions of democracy, even those that share a basic structure with which I agree, need clarification and revision.
Adding to this problem, we are faced with a historical/contextual issue. Practically all definitions of democracy are a distillation of the historical trajectory and the present situation of the originating countries. 14 However, the trajectories and situation of other countries that nowadays may be considered democratic differ considerably from the originating ones. In this case, a theory of adequate scope should assess these differences, per se and as they may generate specific characteristics, and eventually subtypes, of democracies across the whole universe of relevant cases.
Here I argue that current theories of democracy need revision from an analytical, historical/contextual, and legal perspective, even if this entails some loss of parsimony.^15 This effort may yield conceptual instruments appropriate for achieving a better theory of democracy in its various incarnations. The present text attempts to contribute to this task; it is, however, a first step, aimed at clearing conceptual ground. Consequently, in relation to several important topics (especially the relationship of a democratic regime with some characteristics of the state and of the overall social context, as well as with various issues related to the idea of agency) I limit myself to establishing a first connection with these topics; here these connections serve mainly as pointers to themes to be dealt with in future work.
In the coming section, I examine some influential definitions of democracy and draw conclusions that open the way for further analysis.
3. Schumpeter's footnote
After stating that "Democracy is a political method ... a certain type of institutional arrangement for arriving at political - legislative and administrative - decisions", Joseph Schumpeter, 16 offers his famous definition of the "democratic method": "that institutional arrangement for arriving at political decisions in which individuals
(^14) I use this term as a shorthand for referring to the early democratizing countries located in the
15 Northwestern quadrant of the world, plus Australia and New Zealand. For pertinent reflections on the pitfalls premature or unwarranted parsymony, see Keohane, King, 16 and Verba 1994, 20 and^ passim. Schumpeter 1975 [1942], 242.
D emocratic T heory and C omparative P olitics
acquire the power to decide by means of a competitive struggle for the people's vote." This is the paradigmatic "minimalist" (or "procesualist") definition of democracy. However, it is usually forgotten 17 that Schumpeter does not stop here. First, he clarifies that "the kind of competition for leadership which is to define democracy [entails] free competition for a free vote."^18 In the same breath, the author introduces a caveat when, after commenting that "the electoral method is practically the only one available for communities of any size", he adds that this does not exclude other, less than competitive "ways of securing leadership... and we cannot exclude them because if we did we should be left with a completely unrealistic ideal."^19 Significantly, this sentence ends with a footnote that reads "As in the economic field, some restrictions are implicit in the legal and moral principles of the community."^20 The meaning of these assertions, in contrast to the definition Schumpeter offered shortly before, is rather nebulous. The reason is, I surmise, that the author realized that he is close to opening a can of worms: if the "competition for leadership" has something to do with "the legal and moral principles of the community", then his definition or, equivalently, his description of how "the democratic method" works, turns out not to be so minimalist as an isolated reading of the famous definition might indicate.
Furthermore, Schumpeter realizes that, in order for the "free competition for a free vote" to exist, some conditions, external to the electoral process itself, must be met. As he puts it: "If, on principle at least, everyone is free to compete for political leadership by presenting himself to the electorate, this will in most cases though not in all mean a considerable amount of freedom of discussion for all. In particular it will normally mean a considerable amount of freedom of the press."^21 In other words, for the "democratic method" to exist, some basic freedoms, presumably related to "the legal and moral principles of the community", must be effective, and in most cases, as Schumpeter italicizes, "for all". Finally, when this author looks back at his definition and his cognate statement that "the primary function of the electorate [is] to produce a government;"he further clarifies that "I intended to include in this phrase the function of evicting [the government]."^22 Albeit implicitly, Schumpeter makes clear that he is not talking about a one-shot event, but about a way of selecting and evicting governments over time; his definition slips from an event or, as it is often construed, a process - elections - to an enduring regime.
(^17) An exception is Nun 1987, who after noting this omission in the literature (also noted by Held 1987),
18 criticizes Schumpeter for claiming - inconsistently, as we shall see - that his definition is minimalist. 19 Schumpeter 1975, 217; for a similar formulation, 285. 20 Schumpeter 1975, 271; italics in the original. 21 Schumpeter 1975, 271, fn.5. 22 Schumpeter 1975, 271/2; italics in the original. Schumpeter 1975, 272; similarly see 269 and 273.
D emocratic T heory and C omparative P olitics
minimalist, or narrowly centered on the "method", or process, of elections, as its author and most of his commentators took it to be.
Now I will argue that this is also the case, implicitly or explicitly, of all other contemporary definitions that are deemed to be "Schumpeterian", that is to say minimalist and/or "processualist."^26 Among these definitions stands out for its sharpness the one offered by Adam Przeworski: "Democracy is a system in which parties lose elections. There are parties: divisions of interests, values, and opinions. There is competition organized by rules. And there are periodic winners and losers."^27 More recently, Przeworski and collaborators have offered a similar definition, which they label "minimalist": Democracy is "a regime in which governmental offices are filled as a consequence of contested elections. Only if the opposition is allowed to compete, win, and assume office is a regime democratic. To the extent to which it focuses on elections, this is obviously a minimalist definition… [this], in turn, entails three features, ex ante uncertainty,… ex post irreversiblity…, and [repeatability]."^28 Notice that, in spite of its limitation to elections, the irreversibility and, especially, the repeatability of elections in which "the opposition has some chance of winning office as a consequence of elections", 29 imply the existence of additional conditions, à la Schumpeter, for this kind of elections to be held at all. At the very least, if the opposition is to have such a chance, some basic freedoms must also exist.
In his turn, asserting that he is "following in the Schumpeterian tradition" Samuel Huntington defines democracy "[as a political system that exists] to the extent that its most powerful collective decision makers are selected through fair, honest, and periodic elections in which candidates freely compete for votes and which virtually all the adult population is eligible to vote." But this author, adds, as Schumpeter explicitly and Przeworski implicitly do, that democracy "also implies the existence of those civil and political freedoms to speak, publish, assemble, and organize that are necessary to political debate and the conduct of electoral campaigns."^30 Similarly, Giuseppe Di Palma^31 tells us that democracy is "premised ... on free and fair suffrage in a context of civil liberties, on competitive parties, on the selection of alternative
(^26) By this term some authors refer to definitions that purport to focus exclusively in the “process” of elections. Since this meaning is equivalent to “minimalism,” from now on I will use only the latter 27 term when referring to this kind of definition. 28 Przeworski 1991,10. 29 Przeworski et al. 1996, 50/1. Przeworski et al. 1996, 50. More recently, Przeworski (1998) has offered another characterization of democracy in a text that, in spite of its title (“Minimalist Conception of Democracy. A Defense”), 30 moves away from the professed minimalism of the ones I transcribe here. Huntington 1991, 7.
G uillermo O ’ D onnell
candidates for office, and on the presence of political institutions that regulate and guarantee the roles of government and opposition."^32 Larry Diamond, Juan Linz and Seymour M. Lipset offer a more extended but similar definition: "a system of government that meets three essential conditions: meaningful and extensive competition among individuals and organized groups (especially political parties) for all effective positions of governmental power, at regular intervals and excluding the use of force; a ‘highly inclusive’ level of political participation in the selections of leaders and policies, at least through regular and fair elections, such that no major (adult) social group is excluded; and a level of civil and political liberties - freedom of expression, freedom of the press, freedom to form and join organizations - sufficient to ensure the integrity of political competition and participation."^33 On his part, even if Giovanni Sartori centers his attention more on "a system of majority rule limited by minority rights"^34 than on elections, he adds that an "autonomous public opinion ... [and a] polycentric structuring of the media and their competitive interplay" are necessary for democracy to exist. 35 Finally, even though they use a different theoretical perspective, Dietrich Rueschmeyer, Evelyne Huber Stephens, and John Stephens concur: democracy "entails, first, regular, free, and fair elections of representatives with universal and equal suffrage, second, responsibility of the state apparatus to the elected parliament..., and third, the freedoms of expression and association as well as the protection of individual rights against arbitrary state action."^36
Clearly, the definitions I have transcribed are centered on elections of a particular kind, to which they add, in most cases explicitly, some surrounding conditions, stated as freedoms or guarantees that are deemed necessary and/or sufficient for the existence of that kind of elections. Some of these definitions claim to be minimalist à la Schumpeter, but insofar as they must presuppose, at least implicitly, some surrounding freedoms, this claim seems unwarranted. On the other hand, these definitions, whether they claim to be minimalist or not, have the important advantage of being realistic: at least in what refers to elections, they include with reasonable precision attributes whose absence or existence we can assess empirically. To repeat: these definitions do not overlap completely, but all agree on including two kinds of
(^32) Di Palma 1990, 16. (^33) Diamond, Lipset, and Linz 1990, 6/7; italics in the original. (^34) Sartori 1987, 24. (^35) Sartori, 1987, 98 and 110. (^36) Rueschemeyer, Huber Stephens, and Stephens 1992, 43. With the second attribute these authors introduce a new element, which refers to the state, not just to a regime. But this need not occupy us at this moment.
G uillermo O ’ D onnell
the well-being of a collectivity can be viewed as the outcome of a procedure of free and reasoned deliberation among individuals considered as moral and political equals."^42 Again the crucial words are the italicized ones; we are not told in what sense, to what extent, and by whom democracies "can be viewed" as satisfying the requirement stipulated in the definition. Similar objection can be made to the conception of democracy formulated by Jurgen Habermas, as it relies on the existence of an unimpeded deliberative sphere, extremely hard to locate in practice, for characterizing and legitimating democracy and democratic law. 43
Now I invoke another realistic definition, Robert Dahl's polyarchy. 44 I prefer this definition to others of its kind because it is usefully detailed and because the term "polyarchy" allows to differentiate political democracy from other kinds and sites of democracies. This definition shares the structure of the other realistic ones. First, it
(^42) Benhabib 1996, 68; italics added. This definition, as well as other prescriptive ones, omit referring, at least explicitly, to elections. The same is true of some non-prescriptive definitions grounded in rational choice theory, such as Weingast's (1997), where the focus is on limitations on rulers and guarantees of the ruled. Since, whatever the respective author's normative assessment of elections, they are clearly an integral part of existing democracies, this omission seriously hinders the 43 usefulness of these definitions. Habermas 1996, 296: "the central element of the democratic process resides in the procedure of deliberative politics." Habermas (107) adds “Just those action norms [among which are those that “establish a procedure for legitimate lawmaking,” 110] are valid to which all possibly affected persons could agree as participants in rational discourses” (italics added). Niklas Luhmann (1998,
- objects, to my mind decisively, to this and similar definitions: “Every concept of this maxim is explained carefully with the exception of the word ‘could,’ through which Habermas hides the problem. This is a matter of a modal concept, which, in addition, is formulated in the subjunctive. Ever since Kant, one knows that in such cases the statement must be specified by giving the conditions for [its] possibility. That, however, remains unsaid… Who determines, and how does he do so, what could find reasonable agreement?” (italics in the original). John Rawls has recently proposed a definition of legimate law, and by implication of democracy, that is also marred by the problem of proposing hypothetical ideal conditions without stating their conditions of possibility or the consequences of their lack: "Thus when, on a constitutional essential or matter of basic justice, all appropriate governmental officials act from and follow public reason, and when all reasonable citizens think of themselves ideally as if they were legislators following public reason, the legal enactment expressing the opinion of the majority is legitimate law." (Rawls 1996, 770). For balanced assessments of various “deliberative” theories of democracy see Maiz 1996, Johnson 1998 and Fearon 1998. To avoid misunderstandings, I hasten to add that I do believe that deliberation, dialogue, and debate have an important place in democratic politics, and that, in principle, the more there are of these, the better a democracy is. But this does not mean that some idealized or hypothetical public deliberation sphere should be made a definitional component or a requisite of 44 democracy. Of the various, slightly different definitions that Dahl has offered, here I choose the one presented in 1989, 120. Polyarchy consists of the following traits: "1 Elected officials. Control over government decisions about policy is constitutionally vested in elected officials". 2. " Free and fair elections. 3. Elected officials are chosen [and peacefully removed, 233] in frequent and fairly conducted elections in which coercion is comparatively uncommon. 4. Right to run for office [for] practically all adults.
- Freedom of expression. 6. Alternative information , [including that] alternative sources of information exist and are protected by law. 7. Associational autonomy. To achieve their various rights , including those listed above, citizens also have a right to form relatively independent associations or organizations, including independent political parties and interest groups."
D emocratic T heory and C omparative P olitics
stipulates some attributes of elections (clauses 1 to 4). Second, it lists certain freedoms, that Dahl dubs "primary political rights [that] are integral to the democratic process"^45 (clauses 5 to 7),^46 deemed necessary for elections to actually have the stipulated characteristics. At this point I need to stipulate my own definition of elections under a democratic regime.
4. Elections under a Democratic Regime
In a democratic regime, elections are competitive, free, egalitarian, decisive, and inclusive, and those who vote are the same ones who in principle have the right to be elected - they are political citizens. If elections are competitive, individuals face at least six options: vote for party A; vote for party B; do not vote; vote in blank; cast an invalid vote; and adopt some random procedure that determines which of the preceding options is effectuated. Furthermore, the (at least two) competing parties must have a reasonable chance to let their views known to all (potential and actual) voters. In order to be a real choice, the election must also be free, in that citizens are not coerced when making their voting decisions and when voting. In order for the election to be egalitarian, each vote should count equally, and be counted as such without fraud, irrespective of the social position, party affiliation, or other qualifications of each one. 47 Finally, elections must be decisive, in several senses. One, those who turn out to be the winners, attain incumbency of the respective governmental roles. Two, elected officials, based on the authority assigned to these roles, can actually make the binding decisions that a democratic legal/constitutional framework normally authorizes. Three, elected officials end their mandates in the terms and/or under the conditions stipulated by this same framework.
(^45) Dahl 1989, 170. (^46) Slightly rephrasing Dahl, I shall call these freedoms of expression, freedom of (access to alternative)
47 information, and freedom of association. Here I am simply asserting that, at the moment of vote counting, each vote should be computed as one (or, in the case of plural voting, in the same quantity than every other vote). In saying this I am glossing over the complicated problem - which I do not have the space nor the skills to solve here - resulting from rules of vote aggregation that provoke that votes cast in certain districts actually weigh more, and in some cases significantly more, than in other districts (in relation to Latin America and the severe overrepresentation of some districts in some of these countries, see Mainwaring 1999, Samuels and Snyder 1998, and Snyder 1999). Obviously, at some point overrepresentation may become so pronounced that any semblance of voting equality is eliminated, as it happened in some Medieval parliaments, where voting was counted by staments, irrespective of the number of representatives and represented in each statement.
D emocratic T heory and C omparative P olitics
this attribute, arguing that its omission is symptomatic of the degree to which current theories of democracy include unexamined assumptions that should be made explicit for such theories to attain adequate comparative scope. Simply, the literature assumes that once elections are held and winners declared, they take office and govern with the authority and for the periods constitutionally prescribed. 53 This obviously reflects the experience of the originating democracies. But it is not necessarily the case. In several countries there have been candidates who, after having won elections that partake of the attributes already mentioned, were prevented from taking office, often by means of a military coup. Also, during their mandates democratically-elected executives, such as Boris Yeltsin and Alberto Fujimori, unconstitutionally dismissed congress and the top members of the judiciary. Finally, explicitly in cases such as contemporary Chile (and less formally but no less effectively in other Latin American, African and Asian countries) some organizations insulated from the electoral process, usually the armed forces, retain veto powers or "reserved domains"^54 that significantly constraint the authority of elected officials. In all these cases elections are not decisive: they do not generate, or cease to generate, some of the basic consequences they are supposed to entail.
6. On the Components of a Democratic regime, or Polyarchy,
or Political Democracy
Let us remember that realistic definitions of democracy contain two kinds of components. The first consists of assertions of what it takes for elections to be considered (sufficiently) competitive. This is a stipulative definition, 55 not different from "triangle means a plane figure enclosed by three straight lines;" it asserts that elections are to be considered competitive if each one of the attributes spelled out holds. Instead, the second group lists conditions, designated as freedoms, or guarantees, or "primary political rights", that surround fair elections. These freedoms are conditions of existence of an object - competitive elections - to which they stand in a causal relationship. The freedoms complement the stipulative definition with a statement of the kind "In order for X to exist, conditions A…N must exist, too." Similarly as we saw with Schumpeter, as far as I can tell none of the realistic
(^53) Obviously, this possibility is not ignored country and regional studies. The fact that it has barely found echo in democratic theory says a lot, in my view, of the tenacity with which implicit assumptions that have held (and, then, not always correctly) for the originating countries still cling 54 on the contemporary versions of this theory. 55 On Chile, see Garretón 1987 and 1989, and Valenzuela 1992. On definitions in general see Copi and Cohen 1998.
G uillermo O ’ D onnell
definitions make clear if the conditions they proffer are necessary, and/or jointly sufficient, or simply increase the likelihood of competitive elections. This vagueness points to problems I explore below, after noting a third aspect of these definitions.
Above I noted that an assumption of these definitions of democracy, often implicit, is that they do not refer to a one-shot event but to a series of elections that continue into an indefinite future. In saying this we have run into an institution. The elections to which these definitions refer to are institutionalized: practically all actors, political and otherwise, take for granted that competitive elections will continue being held in the indefinite future, at legally preestablished dates (in presidential systems) or according to legally preestablished occasions (in parliamentary systems). This entails that the actors also take for granted that the surrounding freedoms will continue to be effective. In cases where these expectations are widely held, competitive elections are institutionalized. 56 These cases are different, not only from authoritarian ones but also from those where, even if a given election has been competitive, it is not widely expected that similar elections will continue occurring in the future. Only in the first kind of situation relevant agents rationally adjust their strategies to the expectation that competitive elections will continue to be held. Normally, the confluence of these expectations increases the likelihood that such elections will continue happening. 57 Otherwise, elections will not be "the only game in town",^58 and relevant agents will invest in resources other than elections as means to access the highest positions of the regime. 59
This last term needs specification. Slightly modifying the definition Philippe Schmitter and I proposed, 60 by regime I mean the patterns, formal and informal, and explicit or implicit, that determine the channels of access to principal governmental positions, the characteristics of the actors who are admitted and excluded from such access, and the resources and strategies that they are allowed to use for gaining
(^56) For further argument see O'Donnell 1994 and 1996a. (^57) This likelihood of endurance does not mean that after N rounds of such elections a democracy has "consolidated" (as argued, for example, in Huntington 1991), or that other aspects of the regime (as they are deemed to exist in the originating countries) are institutionalized or in the process of becoming so. For discussion of these matters see O'Donnell 1996a and 1996b, as well as the 58 rejoinder by Gunther, Puhle, and Diamandouros 1996. As stated by Przeworski 1991, 26, and Linz and Stepan 1996, 5. Actually, these authors refer not to elections but to democracy as the “only game in town,” but the nuance implied by this difference 59 need not be discussed at this point. Even if agents anticipate that elections at t1 will be competitive, if they believe that there is a significant likelihood that elections at t2 will not be competitive, by a regression well explored in prisoner's dilemmas with fixed number of iterations, agents will make this kind of extra-electoral 60 investments already at t1. O’Donnell and Schmitter 1986, 73, fn.1.