Defeating Executive Assaults in Ecuador

Defeating Executive Assaults in Ecuador: A Coalitional Approach William T. Barndt Department of Politics Princeton University wbarndt@princeton.edu Paper prepared for the Annual Meeting of the Southern Political Science Association, January 2005, New Orleans, LA. DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt South America may be democratic, but it is plagued by presidents who routinely try to restrict basic political freedoms for particular individuals and groups. In this essay, I ask why some of these “executive assaults” succeed, while others fail, by examining the coalitions of actors that form to support and oppose them in Ecuador. In short, I summarize a coalitional approach that identifies (1) which “societal actors” are in conflict over executive assaults in Ecuador, (2) what positions these actors take toward each assault, (3) the coalitions that form between these actors to support and oppose each assault, and (4) which minimal coalitions are necessary and sufficient to produce different assault outcomes. The analysis suggests some disheartening conclusions about the nature of democracy in Ecuador. I. Executive Assaults in South America Political scientists have spilled much ink debating how presidents threaten democracy. Three main arguments can be distinguished. One claims that the separation of legislative and executive power can lead to “deadlock situations.”1A second maintains that presidents are inadequately constrained by the network of state institutions intended to check executive abuses of power.2 Finally, a third analyzes the causes of presidential autogolpes (“self-coups”).3 Each of these contains a measure of truth, but none quite pinpoints one of the most basic problems of presidential rule in South America: Executive Assaults. South Americans presidents usually tolerate opposition voices. But sometimes those voices seem particularly inconvenient: voters threaten to overturn pro-government majorities in elections, street protesters make the president appear indecisive and incapable, newspaper editorials report corruption in the executive branch, etc. Faced with such “inconvenient” opinions, presidents decide Many thanks to the Princeton University Center of International Studies/MacArthur Foundation and the Princeton University Program in Latin American Studies for generously funding the research on which this essay is based. Thanks also to the Universidad San Francisco de Quito and the Center for Development Studies (CENDES) of the Universidad Central de Venezuela for providing institutional affiliations in South America during 2003-2004. 1 Deadlock creates incentives for a variety of different actors to undermine democracy – any of whom may bring down the regime in the end. See, e.g., Linz 1990; Linz & Valenzuela 1994; Mainwaring & Shugart 1997; Shugart & Carey 1992; Carey & Shugart 1998; Cheibub 2002; Cheibub & Limongi 2002.; Pérez-Líñan 2003. 2 O’Donnell led the debate on this issue, beginning with his article on “delegative democracy” (1994). When a president systematically encroaches on the powers of the legislature, courts, and other state agencies, horizontal accountability becomes weaker, making it easier for the executive to rule arbitrarily. Other authors writing about the same cases have also argued that delegative or technocratic styles of presidential rule threaten democracy For example, Conaghan & Malloy (1994) lamented the process by which Andean presidents “usurped” power from other democratic institutions in order to make neo-liberal policy during the 1980s. 3 See, e.g., Cameron (1998). 1 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt that the costs of toleration are higher than the costs of repression (Dahl 1971). Consequently, they order bureaucrats and military personnel – the governing apparatus of the executive branch – to restrict their critics’ freedoms of expression, association, public assembly, or voting.4 In doing so, they initiate what I have called an “executive assault.” (For some examples see Table 1.) Notice that these assaults are not necessarily grand moral battles over Democracy with a capital “D”, but rather more mundane outbursts of presidential frustration. In fact, executive assaults might be best understood as another option in a president’s tactical toolbox. Alongside well-recognized and well-studied means of enacting presidential preferences – like the veto, agenda power, decree authority, etc. – reside less normatively acceptable practices – like bribery and executive assault.5 Table 1: Executive Assaults in South America, Some Examples • • • • • • • Colombia 1981. President Julio Turbay imposes censorship restrictions on news related to the upcoming presidential elections. He orders the dismissal of the director of state television for protesting. Argentina 1986 President Raul Alfonsin suspends union rights following protests in the transportation sector. Brazil 1990. President Fernando Collor de Mello orders the police to imprison store-owners and media figures accused of ‘undermining’ government policy. Peru 1991. In a set of decree laws, President Alberto Fujimori declares heavy penalties for any media that publish government-declared “secret information,” and annuls the results of regional elections. Venezuela 1994. President Rafael Caldera suspends constitutional guarantees enabling the government to arrest suspects without charging them and restrict public assembly. Bolivia 1999. President Hugo Banzer attempts to quash peasant protests by detaining leaders and curtailing freedom of assembly Ecuador 2003. President Lucio Guttierez introduces a new policy that threatens state intervention in response to the publication of insults to the government Observing executive assaults requires a more concrete understanding (operationalization) of the freedoms that presidents try to restrict. I took as my starting point for this question the measurement of democracy more generally. Fortunately, scholars recently turned their attention to This is not to say that all executive bureaucrats and military personnel walk lock-step to presidential orders. To the contrary, throughout most of South America the executive branch is messily divided and contains rebellious, if not openly seditious, components. Yet in the cases I study I find that presidents are almost always able to find some bureaucrats and military personnel willing to carry out the assault. Indeed, this is practically assured by the nature of the patrimonial state discussed in Chapter 2. 5 Executive assault could conceivably be analyzed as an example of what Helmke and others have called “informal institutions.” 4 2 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt evaluating the validity of different measures of democracy.6 Out of this work, consensus has emerged on six unsurprising measures.7 Using these measures, I developed a set of indicators of executive assault. That is, I generated a list of those types of activities undertaken by presidents that seek to restrict the basic political freedoms associated with democracy for particular groups and individuals (see Table 2).8 Table 2: What is an Executive Assault? Accepted Measures of Democracy Indicators of Executive Assault I. II. Regular occurrence of elections Extent of fraud and repression used to secure electoral outcomes9 Extent of the franchise (suffrage) 1. President postpones election polling in specific regions/districts 2. President restricts poll access for certain groups or prominent individuals 3. President directs coercion at the polls at certain groups or prominent individuals 4. President uses electoral fraud systematically to disenfranchise voters 5. President nullifies election results for specific elected officials 6. President arrests, jails, and/or physically abuses journalists for expressing dissident political viewpoints 7. President closes/censors media outlets for expressing dissident political viewpoints 8. President arrests, jails, and/or physically abuses individuals engaged in dissident political protest 9. President bans protests in anticipation of public dissidence 10. President bans/excludes opposition parties are banned from running for office 11. President arrests, jails, and/or physically abuses significant party leaders 12. See also election-related indicators III. IV. Extent of media censorship V. Extent to which dissidents are punished for their political views VI. Extent to which the government controls parties 6 See Bollen & Paxton (2000), Collier and Adcock (1999), Elkins (2000), Adcock and Collier (2001), Munck & Verkuilen (2002), Przeworski et al (2000), Mainwaring et al (2001). See also Studies in Comparative International Development vol. 25, no. 1 (1990) in its entirety, but especially the essay by Coppedge & Reinecke, for the state of the literature a decade ago. 7 This list represents the six indicators that these sources consistently determined were valid. 8 Note that I do not distinguish between those assaults that are constitutionally acceptable (e.g., states of emergency) and those that are extra-constitutional (e.g., sending troops into the streets without congressional approval). I believe the legality/illegality question is less important than we often argue. What matters is the act of assault, not how it is justified. 9 See Lehoucq (2003) on electoral fraud. 3 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Using these indicators, I created a dataset of executive assaults in South American democracies since the onset of the third wave in the region (1979).10 I collected data on executive assaults for the following ten country-periods: Argentina (1983-2002), Bolivia (1982-2002), Brazil (1985-2002) Chile (1990-2002), Colombia (1979-2002), Ecuador (1979-2002), Paraguay (1989-2002), Peru (1980-2002), Uruguay (1985-2002), and Venezuela (1979-2002).11 To collect the data, I turned to Latin American Weekly Report (LAWR) and its associated Regional Reports, which provide weekly summaries of political and economic news in Latin America.12 In short, I decided that an executive assault had occurred when LAWR reported that a president initiated one of the indicators of executive assault listed above.13 A quick glance at the dataset (see Table 3 below) reveals that presidents regularly attempt executive assaults in most South American democracies.14 10 Venezuela and Colombia were democracies well before 1979. Nonetheless, I bound my analysis at 1979 because I believe that the time-period in which democratic breakdown takes place has profound effects on the nature of the breakdown. Bermeo (2003) is good on this point, especially regarding the impacts of the Cuban and Bolshevik revolutions for earlier macro-level breakdowns. 11 One could reasonably question whether all these countries were really democratic across these time periods. More specifically, questions might be raised regarding Brazil (1985-1990), Ecuador (2000), Peru (1993-2000), Venezuela (1999-2002) and Paraguay (1989-1991) Indeed, Carothers (2002) recently observed that many polities inhabit a “gray zone” in which neither authoritarian nor democratic regimes are institutionalized. Similarly, Diamond (2002) has written, “Interesting issues revolve around the boundaries between regime types, which [are] blurry and controversial. When fitting messy and elusive realities against ideal types, it cannot be otherwise” (27). But fitting ideal types to a messy reality is not what hinders classifying polities as democratic or authoritarian. Instead, it is the fact that so many polities sit on the tipping point between democracy and authoritarianism. Such polities barely meet the criteria for democracy, fall just below the minimum, or fluctuate around that point. Nonetheless, their proximity to minimal democratic standards led me to include them in my sample. In fact, I would argue that attempts to pigeonhole these polities into regime types do not make much sense – I am skeptical of the usefulness of the hybrid-regimes category in Latin America outside of a very small number of cases (like PRI-ista Mexico). 12 LAWR is the most comprehensive and reliable weekly summary of news on all of the Latin American countries. Moreover, it harbors much less bias towards the larger, developed countries in LAWR than other similar publications. Nonetheless, using LAWR as my source of data probably biased my data in notable, but surmountable, ways. First, I suspect that LAWR mostly reports the highly politicized executive assaults, ignoring lower-profile assaults. Moreover, LAWR no doubt underreports those assaults that never hit the headlines in the countries where they take place. My dataset thus contains information on publicly contested executive assaults—not all executive assaults. Second, I probably mis-classify as executive assaults some simple cases of political posturing; that is, instances when micro-level democratic breakdown is not really at stake. LAWR articles are occasionally unclear whether or not the executive is really undertaking an assault or whether opposition groups are simply spinning less pernicious presidential behaviors for their own political advantage. Attempting to correct for this bias, I rely on secondary sources to help decide unclear cases (I also slightly modified the Ecuadorian and Venezuelan data based on my fieldwork). 13 Observing attempted executive assaults was a sometimes tricky enterprise—the president often attempts micro-level democratic breakdown in coalition with other significant actors (most often other branches of the state). As such, I only recorded those instances of micro-level breakdown when it was apparent that the president was the key initiator. In practice, this often meant that the president was linked to the attempted breakdown with an “active verb” (e.g., ordered, declared, decreed, tried, etc.). 14 There I have come to the discouraging conclusion that it is unlikely that executive assaults in most South American democracies will diminish; that assaults are a policy likely to be attempted just as frequently in the future as they have been thus far. In short, I argue that executive assaults represent moments in which tension peaks between the two mutually incompatible ‘logics of governing’ that coexist in neopatrimonial democracies. In these moments, Dahl’s costs of repression fall below the costs of toleration. 4 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Table 3: Executive Assaults in Third-Wave South America* Country Ecuador Period Analyzed % Country-years with at least one executive assault 58% 1979-2002 (24 yrs) Venezuela 1979-2002 42% (24 yrs) Brazil 1985-2002 38% (18 yrs) Argentina 1983-2002 30% (20 yrs) Colombia 1979-2002 29% (24 yrs) Bolivia 1982-2002 28% (21 yrs) Paraguay 1989-2002 28% (14 yrs) Peru 1980-2002 26% (23 yrs) Uruguay 1985-2002 11% (18 yrs) Chile 1990-2002 0% (13 yrs) *These figures exclude assaults that respond directly to guerrilla attacks Presidents routinely attempt executive assaults, but not all their attempts succeed. Opposition voices are not silenced by simple presidential pronouncement alone. Despite all efforts, assaulted individuals and groups are sometimes able to ward off the executive. This begs the question: what is the difference between a successful and a failed assault? How can an assault be defeated? Assaults succeed when a president prevents the individuals (at whom the assault is directed) from expressing their “inconvenient” opinion in the way they choose. Assaults fail when presidents are unable to prevent them from doing so.15 Coding assault outcomes requires three pieces of information: (1) who is assaulted; (2) the inconvenient opinion that is at stake; and (3) the way that opinion is expressed. With this information, it is usually relatively simple to decide if the assaulted individuals or groups could express themselves in spite of the president’s attempt to silence them. I grant that there are marginal cases that can arise when coding assault outcomes. Imagine, for example, that the president issued a decree suspending a political party of the Left. Protests resulted during which several union leaders were jailed. Nonetheless, the protests convinced the president to revoke his decree…but the union leaders remained imprisoned. Would such a case be a success or a failure? In fact, I would code it as two different assaults, one which succeeded and one which failed. Yet this was not a real problem, as I observed no such outcomes in the cases I analyze here. In the end, all I can do for skeptics encourage interested readers to return with my coding rules to the same sources I used to collect my information. 15 5 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Ecuador provides a useful window into the question of assault outcomes. It was the first Latin American case of third-wave re-democratization, thus has a twenty-five year history of (moreor-less) uninterrupted democratic rule during which eleven presidents have governed (seven of whom were elected). It has not only experienced a large number of publicly contested executive assaults (eighteen), but also significant variation in the outcome of those assaults (ten succeeded, eight failed). In Ecuador, assaults usually assume one of four forms: repressing popular protests; jailing elites; muzzling radio stations and the press; or manipulating congressional elections. These assaults tend to occur in response to inconvenient opinions on one of three subjects: the president’s capacity to manage the economy; the president’s efforts to ensure his allies prominent places in the state bureaucracy; or the president’s integrity/competence. Table 4 (below) summarizes the eighteen executive assaults I observe in Ecuador.16 II. A Coalitional Approach to Assault Outcomes in Ecuador A long tradition in the social sciences explains political outcomes by analyzing the coalitions that form to support and oppose them.17 The logic of coalitional approaches is straightforward. Different actors have different preferences for different outcomes. Given a range of possible outcomes, which comes to pass depends on the relative strength of the opposing coalitions that support each one. A coalitional approach seems quite appropriate for the analysis of assault outcomes. During executive assaults, I observe two sides initially in conflict with one another:18 the president, who wants the assault to succeed, and the assaulted individuals, who are trying to defeat the president’s attempt to restrict their political voice. The conflict thus consists of each side trying to convince other groups to support or oppose the assault. Different groups have different resources at their disposal, some more powerful than others. Whether an assault succeeds or fails depends on which of the opposing sides brings the most powerful combination of resources to bear on the conflict: that is, on the relative strength of opposing coalitions. Conflicts over executive assaults are extraordinarily messy: scores of individuals and associations take positions on most assaults. To be analyzed effectively, the vast number of people involved in any given conflict must be reduced to a manageable set of actors. How to do this is at These are drawn directly from my dataset (described above). Coalitional approaches have a long history in the social sciences (e.g., Marx 1852, Moore 1966, RSS 1992, Luebbert, Berg-Schlosser and Mitchell 2002, Alexander 2002, Przeworski 1991, Geddes 1991, Gibson 1997, Ames 1987, O’Donnell & Schmitter 1986, Gourevitch 1985, Frieden 1994, Katzenstein 1986, Rogowski, Bates 1981, Karl 1997). 18 Note that this is an assumption, rather than a statement of fact. 16 17 6 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Table 4: Executive Assaults in Ecuador (1979-2003) ***Cases of Failed Assaults are shaded gray; Cases of Successful Assaults are not shaded*** Date Mar-May 1980 October 1982 President Roldós Hurtado Inconvenient Opinion Expressed Opposition in Congress to drafted plan that would allow president to dismiss the legislature and rule by decree. Opposition in streets and in media to President’s economic adjustment program, including changes to minimum wage; job security; public transport fares; price of basic foodstuffs/services; and debt payment Opposition in Congress and in media to President’s economic restructuring program and his attempt to stack Supreme Court Opposition in streets to President’s economic restructuring program Widespread and general opposition to President provokes attempt to augment executive state-of-emergency powers Opposition in streets to President’s economic restructuring program, including hikes in prices of basic foodstuffs/ services and public transport fares; and debt payment Opposition in streets, in media, and in Congress to President’s economic restructuring program and his condoning of police abuses under the Minister of Interior Opposition in streets to President’s minimumwage freeze and hikes in public transport fares Possibility of corrupt dealings between the military and President’s brother (on radio) Opposition desire to vote against President’s congressional coalition in mid-term elections Opposition in streets to President’s agrarian reform Opposition in streets to President’s fuel pricehikes, privatization plans, and wage freeze Opposition in streets to President’s plan to ‘modernize’ urban transport; demands for wage increase Refusal to broadcast President’s personal talk show on radio, if he uses it to “hurl insults at the opposition” Opposition in streets to President’s economic restructuring program, including freezing bank accounts and minimum wage; hikes in fuel prices and public transport fares; privatization; and social-security reforms Public statements critiquing President’s ability to manage economy effectively Opposition in streets to President’s hikes in fuel prices and public transport fares Press critique to President’s competence Assaulted Groups/Individuals (More specific or leave general?) Congressional opposition, including dissident CFP and ID deputies Workers; Teachers; Radio Broadcasters; Students; Transportistas Radio Broadcasters; Congressional deputies Political Left, Labor, Students Protesters; Voters Workers Relation to Indicators 5 8 10 7 8 9 5 6 7 8 8 9 8 9 6 7 8 11 8 7 8 1 12 8 9 8 9 8 9 7 8 9 Outcome Assault Fails Assault Succeeds Oct-Nov 1984 December 1985 January 1986 September 1986 Sept-Dec 1987 June 1988 Aug-Sept 1990 Mar-Aug 1994 June-July 1994 May-June 1995 Jan-Mar 1996 January 1997 June-July 1999 Febres Cordero Febres Cordero Febres Cordero Febres Cordero Febres Cordero Febres Cordero Borja DuránBallén DuránBallén DuránBallén DuránBallén Bucaram Mahuad Assault Succeeds Assault Succeeds Assault Fails Assault Succeeds Assault Succeeds Assault Succeeds Assault Succeeds Assault Fails Assault Fails Assault Succeeds Assault Succeeds Assault Fails Assault Fails Congressional deputies Radio Broadcasters; Workers; Teachers; Students Workers; Students; Teachers Radio Sucre, a station owned by an opposition congressman Registered Voters Indigenous; Workers; Students; Transportistas Students; Transportistas; Workers; Indigenous Workers; Indigenous; Transportistas Members of Ecuadorian Association of Radiobroadcasters (AER) Indigenous; Workers; Transportistas; Students August 2000 Jan-Feb 2001 July-Nov 2003 G. Noboa G. Noboa Gutiérrez President of National Federation of Exporters Students; Workers; Indigenous; Transportistas Journalists (UNP, FENAPE); Newspaper Owners (AEDEP) 8 8 9 6 8 Assault Fails Assault Succeeds Assault Fails 7 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt the heart of designing a coalitional approach. Should we analyze coalitions between particular associations (e.g., particular unions, branches of the military, and media outlets)? Or should we aggregate these associations into more abstract “societal actors”19 like the Working Class, the Armed Forces, and the Media? Which approach makes the most sense for the conflicts at hand? Here I chose to try to aggregate at a fairly high level, for a number of reasons.20 Of these, the most theoretically interesting was to explore the validity of the extensive body of work on regime change that is populated by abstract actors like “The Bourgeoisie” and “The Working Class” (e.g., Rueschemeyer et al 1992; Moore 1966; Luebbert). If this approach is anything more than socialscience fantasy, we should be able to observe the same types of actors in more quotidian conflicts over the practice of democracy. I understand these abstract or “societal” actors to be sets of associations loosely tied together by similar ideas and interests. Just as political economists see firms and unions as the building blocks of sectors or classes, I see civil associations as the building blocks of the abstract “societal” actors analyzed here. This loose understanding is purposeful – I do not wish to restrict the types of actors I analyze to Marxist classes, Weberian status-groups, or the sectors of modern political economy. But such openness is a double-edged sword: how then to decide which actors to include? In fact, this is much of what model specification means for coalitional approaches. For this analysis, my inclusion rule was rather simple: I included all those societal actors that existing scholarship argues are theoretically relevant for conflict over democratization21 and that The term is awkward, though Gourevitch (1985) uses it well. Why do I choose to analyze coalitions between societal actors instead of coalitions between individual associations? First, over the twenty-five year period I analyze individual associations rise and fall, but the societal actors to which they belong have a certain degree of continuity. For example, over the past fifteen years the associational membership of the property-less in Ecuador has shifted from working-class unions to indigenous associations (see Yashar forthcoming on this point). Working-class unions have little importance today, compared to what they had in the early 1980s (and vice versa for indigenous associations). But because I consider both types of associations as members of “The Property-less,” I am able to analyze the effects of The Property-less across the full twenty-five year period (more on this specific case below). Second, the same associations do not always take positions on executive assaults. By aggregating associations into societal actors, however, I can make claims about “The Armed Forces” supporting two different assaults even if it is the Army supporting the first and the National Police supporting the second. In short, working with societal actors allows me to simplify my analysis and representation of the many complex conflicts that take place over assaults. Reducing assault conflicts to a handful of aggregate actors makes a coalitional approach to a larger number of assault conflicts feasible. Moreover, using societal actors opens up the possibility of generalizability beyond the immediate contexts. “Capital” makes sense in many countries (though by no means all); the Quito Chamber of Commerce has meaning only in Ecuador. 21 To identify theoretically relevant actors, I conducted a review of the literature on democratization in South America. This produced a rather long list of actors Capital, Landlords, The Urban Working Class, The Middle Class, Peasants, Indigenous Groups, The Armed Forces, State Bureaucrats, The Law, The Catholic Church, Guerrilla Groups, New Social Movements, Human Rights Groups, International Actors (including states, financial institutions, and NGOs), The Media, Universities, and Political Parties. 19 20 8 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt actually take form during assault conflicts in Ecuador.22 This approach produced six actors: Capital, The Property-less, The Armed Forces, The Law, Human Rights Networks, and The Media. The inclusion of Capital and The Property-less confirms the continued significance of socioeconomic class for conflicts over democracy. Nonetheless, conflict over executive assaults cannot be reduced to class: components of the state – even the weak and uneven Ecuadorian state – are also in conflict over democracy. Showing up repeatedly are The Armed Forces, whose importance for sustaining or undermining democracy has a long history.23 Also present are the cluster of legal institutions – which I personify as The Law – that have been subject to a recent rash of work on horizontal accountability and the “rule-of-law” (O’Donnell, etc.). Beyond state and class, Human Rights Networks have become regular participants in conflicts over democracy (see, e.g., Keck & Sikkink). And, finally, the independent importance of The Media in developing democracies has just begun to be recognized (See, e.g., Boas; Lawson). The types of associations I included in each actor are listed in Table 5 below; the lists of specific interest groups and associations are in an Appendix.24 22 Each assault was analyzed in the following manner. I began by reading journalistic accounts of the assault in El Comercio (Quito), El Universo (Guayaquil), El Hoy (Quito), &/or El Telegrafo (Guayaquil). By tracing the process by which the assault unfolded, I developed a more complete understanding of the assaults I had identified in the dataset. Moreover, it allowed me to identify the public stances taken by different state and social actors on the executive assault and the policy to which it related in two ways. First, representatives of different organizations interested in the assault outcome were regularly quoted as the assault unfolded. Second, Ecuadorian organizations regularly publish public statements of their political positions in the major newspapers. Using these two types of information, I was able to identify which organizations were publicly pressuring the president and the positions those organizations adopted. Where necessary and possible, I conducted interviews with the actors who were involved in the conflict to clarify the positions they took during conflict. For each case, I thus developed a list of which organizations made public statements about the assault, the position they took, and the intensity of that position. In this way, I identified more than eighty Ecuadorian organizations that took public positions on one or more of these executive assaults. The types of organizations included in each grand political actor are listed in Table below; the specific Ecuadorian organizations included in each are listed in Table (attached). See also Appendix (attached). 23 See, e.g., See Stepan 1973; Lowenthal & Fitch 1986; Fitch 1998; Nun 1976; O’Donnell 1973/1978; and Zimmerman 1983. 24 Let me add a quick note on some actors not included in the analysis. First of all, the Middle Class is currently a non-actor in most of Latin America having been splintered and decimated by two decades of economic depression. Second, Landlords are irrelevant having lost their peasants, been abandoned by the Church, and displaced by capitalist agriculture [RSS comment on this in their Conclusions]. Third, although Peasants, Indigenous Associations, and the Urban Working Class have brought diverse identities and claims to the political arena, I do not see relevant distinctions between them in terms of the resources they use to contest assaults. As, such I lump them all together in The Property-less. Fourth, the Catholic Church is still an active player in these conflicts but is included as part of Human Rights Networks, given the long-recognized seismic shifts in its positions towards human rights abuses. Indeed, the Church-PeasantLandlord trio once called “Conservatism” fractured and disintegrated long ago in most of South America (yet for some reason still pops up in our analyses). Fifth, International Actors are clearly important participants in some conflicts, but do not form a coherent societal actor as such. Instead, I include individual international organizations as members of other societal actors (see table below). Finally, and perhaps most controversially, I see Political Parties as having little importance for assault conflict in most of South America. Indeed, their importance approaches zero in Ecuador, where the loose associations of elites we choose to call parties are barely recognizable as such. Levitsky’s 1999 comment about Peruvian parties under Fujimori as “electoral devices” is dead on, and just as applicable to Ecuador. 9 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Table 5: Ecuadorian Actors Contesting Executive Assaults Capital 1. 2. 3. 4. 5. Commercial sector Industrialists Capitalist Agriculture Financial sector Information sector The Property-less 1. Working Class unions 2. Indigenous associations & Peasant unions 3. Informal sector Human Rights Networks 1. Domestic and international human rights NGOs 2. Election observation teams 3. Catholic Church 4. Students 5. Universities 6. Artistic/Intellectual community The Media 1. Media owners and directors 2. Journalists 3. Broadcasters 4. International media organizations The Armed Forces 1. Military Branches 2. Police Forces The Law 1. Supreme Court 2. Electoral Court 3. Constitutional Court 4. Congress 5. Defensoría del Pueblo 6. Mayor of Quito 7. Lawyers’ Associations ***For a listing of the specific Ecuadorian organizations contained in each actor, see Appendix*** It is important to remember that the conflicts I portray as taking place between societal actors are in fact between the melee of associations that compose society (see Table attached). Abstract actors like “The Property-less” do not ‘really exist’: they are simply useful representations of those moments when loosely related associations take the same position, bringing their resources to bear on the conflict. Yet the associations that compose an actor do not always come together around the same position during each assault conflict. If the major associations that compose an actor are bickering among themselves over the assault (or could care less about it), that actor cannot take a position towards the assault.25 For example, if Ecuadorian industrialists, commercial farmers, and financiers disagree over whether an executive assault is justified, it would be ridiculous to claim that Capital had taken a defined position regarding the assault.26 In such instances, I instead assume that the actor simply does not exist – and thus cannot reasonably be assigned a role as a protagonist of the conflict. When I say that an actor takes a position, therefore, I mean that an important portion of the associations that compose it are taking similar positions towards the assault (perhaps coordinated, but usually not) and no major associations are taking contrary positions.27 For each assault, I thus recorded the position taken by each actor as either Supports the assault, Opposes the assault, Though it is still feasible that particular associations that make it up might be, a point I take up below. Again, more specific coding rules for when actors take positions are listed in the Appendix. 27 On the problem of corporate agency, see Wendt 1999 and recent work by Philip Petit. 25 26 10 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Neutral, Divided, or Uninvolved (more detailed coding rules are listed in Table 11 in the attached Appendix). What did I expect to find these actors doing during assault conflicts? Which would be fighting to defeat assaults? Which would be supporting the president? Which coalitions would be associated with failed assaults? Which would be associated with successful assaults? In fact, I had few prior beliefs about the answers to these questions. I thought I would find each of the six actors supporting some assaults and opposing others. I expected these shifting alliances to produce a handful of coalitions associated with successful assaults and a handful of coalitions associated with failed assaults. In short, I expected reality to be fairly complicated, with no one actor or combination of actors somehow ‘key’ to determining assault outcomes. But reality surprised me: as will be seen, the results of this analysis demonstrate these expectations to be profoundly mistaken. So what (if any) positions did each of the six actors take towards the eighteen executive assaults I observe in Ecuador? Table 6 below lists the percentage and number of times that each actor adopted each position. The final two columns indicate how often the associations that compose each actor cohere and take a position on an assault. Table 6: Frequency with which Societal Actors Emerge During Assaults in Ecuador Support Armed Forces Capital Property-Less The Law Human Rights Networks The Media a Oppose 0% 0 0% 0 72% 13 44% 8 72% 13 22% 4 Neutral 11% 2 6% 1 6% 1 0% 0 0% 0 0% 0 Divided 11% 2 33% 6 11% 2 33% 6 6% 1 28% 5 Uninvolved 22% 4 33% 6 11% 2 22% 4 22% 4 50% 9 56%a 10 28% 5 0% 0 0% 0 0% 0 0% 0 Emerges as Societal Actor (Support, Oppose, or Neutral) 67% 12 33% 6 78% 14 44% 8 72% 13 22% 4 Does Not Emerge as Societal Actor (Divided or Uninvolved) 33% 6 67% 12 22% 4 56% 10 28% 5 78% 14 This indicates that the major organizations that make up The Armed Forces gave their support to 56% of the executive assaults I observe in Ecuador (10 out of 18 assaults) [Coding rules for each grand actor are appended] An important fact emerges out of this table (the first indication that my initial expectations might be mistaken): in Ecuador, societal actors do not vacillate between supporting and opposing 11 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt assaults. Neither the Armed Forces nor Capital ever openly opposed an executive assault, which is to say (1) the associations that comprise the Armed Forces never united together against an assault and (2) the associations that comprise Capital never united together against an assault. Both either supported the assault, on one hand, or did not support the assault (were divided, neutral, or uninvolved) on the other. Conversely, none of the other grand actors (The Property-less, The Law, The Media, or Human Rights Networks) ever gave its united support to an executive assault. Each either opposed the assault, on one hand, or did not (was divided, neutral, or uninvolved) on the other. The consistency of these positions is remarkable: in Ecuador it seems that abstract actors can be portrayed as if they had ‘regime preferences’, an idea I tend to view with skepticism. But it also demonstrates why we cannot simply assume that an actor will take a particular position toward an executive assault based on this knowledge. Societal actors may not vacillate between supporting and opposing assaults, but (as discussed above) the expression of their ‘preferences’ is contingent on the associations that compose them adopting similar positions towards an assault. Since this does not happen in every conflict, the expression of an ‘actor’s preference’ during a given conflict must be empirically verified before we assume it plays a role in determining the assault outcome. These findings have important operational consequences for the analysis. First, given that the Armed Forces and Business never take unambiguous positions against an executive assault, I assume that what matters is whether they support the president, on one hand, or do not support the president on the other. That is, the relevant variation in the position taken by these two actors is Support vs. Non-support for the assault (i.e., whether they did or did not bring the resources they control to bear on the assault).28 Second, given that none of the other four actors ever take a united position in favor of an executive assault, I assume that what matters for the outcome is whether they oppose the president, on one hand, or do not oppose the president on the other. That, is the relevant variation in the position taken by these four actors is Opposition vs. Non-opposition to the assault (i.e., whether they did or did not bring their resources to bear on the assault). I have thus simplified complex conflicts over executive assaults in Ecuador to the positions taken by six “societal” actors. I have created, in other words, a model consisting of six independent variables, measured dichotomously: Capital Support for an assault (Yes/No); Armed Forces Support for an assault (Yes/No); Property-less Opposition to an assault (Yes/No); Law Opposition to an 28 I assume that divided, neutral, and uninvolved have equivalent affects on the outcome. This, of course, is not entirely accurate since the non-opposition or non-support of an actor can mean either that the resources that actor controls are not brought to bear on the resource or that partial resources are rallied to one or both side. But it is a useful simplifying assumption. 12 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Table 7: Assault Conflicts in Ecuador Armed Forces Business Property-less State Legal Human Rights The Media Assault Support Supports Oppose Bodies Oppose Networks Oppose Opposes Outcome? Assault? Assault ? Assault? Assault? Assault? Assault? 1a Yes Yes Yes No Yes No Succeeds 2 Yes Yes Yes No Yes No Succeeds 3 No No Yes Yes No No Fails 4 No No Yes Yes No No Fails 5 No No Yes No Yes No Fails 6 No No Yes No Yes No Fails 7 No No No Yes No No Fails 8 Yes Yes Yes No No No Succeeds 9 Yes No Yes Yes No No Succeeds 10 Yes No No Yes No No Succeeds 11 Yes No Yes No No No Succeeds 12 No No No Yes Yes No Fails 13 No No No No No Yes Fails 14 Yes No Yes Yes No Yes Succeeds 15 Yes No No No Yes Yes Succeeds 16 No No Yes No Yes Yes Fails 17 Yes Yes Yes No Yes Yes Succeeds 18 Yes No Yes Yes Yes Yes Succeeds a Row 1 shows that the Armed Forces and Business supported the executive assault, while the Property-less and Human Rights Networks mobilized against it. Neither The Law nor The Media cohered as actors during the assault. Assault Number assault (Yes/No); Human Rights Networks Opposition to an assault (Yes/No); and Media Opposition to an assault (Yes/No). In effect, these are the six causes that I believe to be most relevant to whether an executive assault in Ecuador is resolved in favor of the president or the assaulted individuals. [The observations for each of the eighteen assaults are reproduced in Table 7 below].29 Each row in Table 7 represents the opposing coalitions that came together during an assault conflict; that is, it shows which actors supported each assault and which actors opposed it. More broadly, I have shown how societal actors can be used to analyze smaller-scale conflicts over democracy, if existing theory about actor relevance is combined with inductive research about which actors actually exist, when they exist, and the positions they take in those moments.30 Three rows are repeated twice, the other twelve are unique. An important issue still must be resolved: are there factors other than the positions assumed by grand actors that might influence the outcome of an executive assault? First, it is rightly observed that contextual factors – like crises or electoral cycles – tend to shape and change actor preferences. Yet I am not (yet) interested in the causes of actor preferences, only in their effects. As such, I do not consider any alternative explanations whose effects are realized through the position taken by societal actors – as most contextual factors are. In this sense, my analysis has a clear bias towards extremely proximate causes. Second, many coalitional approaches to policy-making maintain that the effects of actor preferences shift as they are voiced to state officials through institutions and systems of interest intermediation. To understand the effects of coalitions, they argue, we must examine how preferences are articulated to the state. While this may be true in some cases, it does not seem particularly relevant to Ecuador, which lacks an institutionalized system of interest intermediation and where institutions (“rules”) are deeply politicized. As such, I do not see either of these variables intervening decisively between preferences and outcomes [continued on following page]…. 29 30 13 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Having reduced executive-assault conflict in Ecuador to a quite parsimonious form, I now introduce a crucial complication: for any given assault, the positions taken by the six societal actors tend to be deeply interconnected, in two different ways. First, the effect of the position of any one actor – say, the opposition of Law – on whether an executive assault succeeds or fails may well depend on the positions of other actors. For example, whether or not the opposition of the Law can defeat an assault may depend on the positions taken by other actors. The logic here is that the effect of the same value of the same independent variable on the dependent variable may change, when the values of the other already specified independent variables change). Second, the position taken by an actor during an assault is rarely independent of the positions of the other actors. Thus far actor-positions have been described as if they were separable components of conflicts over assaults: “The Armed Forces support the assault” or “Capital is divided.” This imagery is reinforced in Table 7, which presents the final position of each actor towards each assault. In fact, the final position of an actor is usually defined over the course of a conflict (see below). During the conflict, the associations that comprise each actor are aware of and often influenced by the positions taken by the associations that make up other actors. As such, the final position of each actor is usually intertwined with the positions of the other actors. Once this is recognized, however, it makes little sense to argue inflexibly that each actor-position is a separable part of the assault conflict. Slicing up each conflict into separable, analytically distinct, “independent” pieces misunderstands and misrepresents executive assaults. Given that actor positions are interrelated, at least initially it makes more sense to There is, however, a different and quite relevant way to think about alternative explanations: Does the support or opposition of an actor have the same effect on assault outcomes for all eighteen Ecuadorian assaults? (See Ragin 2000 on this question.) Is it not possible that the same actor-position has different effects on assault outcomes at different times? That is, the effect of the same value of the same independent variable on the dependent variable may change, when the values of other unspecified independent variables change. This might happen in three ways. First, short-run contextual factors might change the effects of an actor’s position on assault outcomes. For example, the support of capital for an assault might be more important during economic booms (when capital has more resources at its disposal), and less important during economic crises. At least in Ecuador, however, this is not seem the case: short-run contextual changes seem to impact the position an actor takes, not the meaning of that position for the outcome. Second, long-run changes in actor “strength” could change the meaning of a given actor position over time. For example, the opposition of the Property-less in 1980 may not mean the same thing as the opposition of the Property-less in 2003? Though there have been some shifts in the relative power and composition of Ecuadorian societal actors over the past twenty-five years – for example, the virtual replacement of the labor-syndicalist movement with the indigenous movement – they do not substantially affect the results reported below. Finally, actor positions may have different effects on assault outcomes in different presidencies. The opposition of the Law may well have different effects on President Febres Cordero (who tried to disband the Supreme Court) than the opposition of the Law to President Borja (who is widely viewed as having respected other branches of government.). This would be a serious concern with the design, were it not for the fact that the results I uncover below demonstrate that it is not the case. 14 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt assume that each different combination of causal factors (i.e., each unique row in the Table 7 dataset) is a “package” (Ragin 2000: 72) or a “bounded whole.”31 These complications are necessary in order to analyze assault conflicts without profoundly misrepresenting them. Yet they present a serious headache for mainstream analytic techniques in political science, like regression or qualitative analysis in the style of King, Keohane, Verba (1994). As Ragin (2000) and others have noted, such techniques usually assume (1) that outcomes can be predicted by summing the weighted scores of all independent variables in the model and (2) that the score on any given independent variable will have the same effect on the outcome regardless of the scores taken by the other independent variables.32 Neither of these assumptions is reasonable given the nature of assault conflicts. Fortunately, methods are available – most thoroughly developed in the work of Charles Ragin (1987, 2000) – that allow us to deal with a larger number of causally complex cases without imposing untenable assumptions on the social world. 33 Ragin embraces the premise that each unique combination of independent variables (unique row in Table 7 above) is a potentially unique type of cause. Moreover, instead of assuming that the summed effect of the six “independent” variables determines assault outcomes, Ragin (initially) assumes it is the effect of the interaction between all six causes that determines assault outcomes. In short, a clear congruence exists between the assumptions that underlie Ragin’s understanding of causal complexity and my understanding of the nature of executive-assault conflict. As such, these techniques are peculiarly well-suited to analyzing assault conflicts.34 31 Collier & Adcock (2001) and Sartori (1984) use this term to refer to how we think of concepts. I find it just as useful for thinking about cases. 32 Note that this is the logic of both case studies and interaction effects in mainstream statistical techniques 33 Thus addressing the problems of two other methodological options. First, case-study research and processtracing does this by presenting carefully detailed stories about specific instances of the phenomena being analyzed. For practical reasons, such analyses tend to deal with a small number of cases. Yet here I am comparing eighteen different assault conflicts. A paper that traced the processes in eighteen cases would be unbearably lengthy and repetitive for the reader, not to mention time-consuming for the author. For practical reasons, therefore, case-study research seems inappropriate for this analysis. Second, game theory explicitly models interactions between different actors. Yet I do not believe that the interconnection between abstract “societal” actors should be modeled as “strategic interaction among forward-looking agents” (Frieden 1991: 40), given the fact that they are not so much taking decisions as reflecting the decentralized choices of disparate organizations. Ecuadorians are not thinking in terms of the reactions of abstract actors when they make their decisions, I am simply portraying the conflicts that way. The societal actors analyzed here are not six poker players sitting around a table, they are abstractions I have created to help simplify and make sense of assault conflict in Ecuador. (One could, I suppose, develop a game-theoretic approach to when grand actors actually congeal. But this would be a discussion about why grand actors exist in some moments and not others which, for the moment, is prior to the analysis.) 34 I would argue that choices of analytic techniques should not be based on ontological preferences, ideological biases, or sub-disciplinary divisions. Instead, we should carefully consider the extent to which the assumptions demanded by different analytic techniques correspond to the reality of observed events. 15 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt From this foundation, Ragin has developed techniques that use the logic of necessity and sufficiency to identify the least complex combinations of causes identified with different outcomes.35 Basically, the analysis works as follows: Each combination of causes that led to the same outcome is compared with every other combination that led to that outcome. All causes that appear logically superfluous to an outcome are eliminated. Ragin & Gisele explain: “more-complex expressions can be absorbed by less-complex expressions as long as all the elements that appear in the less-complex expressions also appear in the more-complex expression” (2002: 75; see also Ragin 2000: Chapters 4 & 9). In other words, the analysis identifies the least complex combinations of actor-positions associated with successful assaults and the least complex combinations of actor-positions associated with failed assaults. These remaining “least-complex” or “minimal” combinations of causes are considered to be sufficient to cause the outcome in question. And this is the key: These leastcomplex combinations represent the minimal coalitions that lead to either assault success or assault failure. 36 Any causes common to all the minimal combinations that led the same outcome are considered to be necessary causes of assault success (or failure).37 III. The Power of Guns and Money This analysis produced remarkably clear results. Contrary to my expectations, the diverse conflicts I observed in Table 7 (above) can be reduced to rather simple coalitions. It turns out here is only one pathway to assault success and one pathway to assault failure in Ecuador (see Table 8 below). To begin, the support of the Armed Forces was the sole necessary and sufficient cause of successful assaults. In all instances when the Armed Forces supported the president, the assault succeeded. In all instances when the assault succeeded, the Armed Forces supported the president. In other words, we can claim (1) once the Armed Forces support the assault, it will succeed; (2) Armed Forces support is the only way assaults succeed; and (3) the positions adopted by the other five actors do not seem to have an effect on assault success.38 Given the way Armed Forces support 35 On necessity and sufficiency see Braumoeller & Goertz 2000; Goertz & Starr 2002; Ragin 2000; Goertz & Mahoney 2004. Ragin’s software (QCA) is available at http://www.u.arizona.edu/~cragin/fsqca.htm. 36 No coalition led to both assault success and failure. 37 Indeed, I would argue that the above discussion gives good reason for why Ragin’s methods are particularly well-suited to coalitional analyses more generally. 38 How confident can I be in this result? It bears noting that I observed no inconsistent cases: all ten cases of successful assaults were characterized by the same cause. Yet we are only examining ten cases of successful assaults. To establish the veracity of my results, I use a binomial probability test to ask ‘What is the probability of observing ten successes out of ten trials, if the underlying probability of an assault succeeding is 65%?’ (this language paraphrases 16 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt is coded, this means that at least one branch of the armed forces is supporting the assault through public statements and/or direct repression and no branch was fomenting dissent, either openly or behind the scenes. The only meaningful coalitional partner for the president, it turns out, is the Armed Forces. It may come as no surprise that successful assaults always entail a certain armed presence, be it for repressing protests, jailing or beating up individuals, or confiscating property. More surprising, however, is the fact that Armed Forces support is the only factor consistently related with successful assaults. Second, the various conflicts associated with failed assaults reduce to two necessary causes that are jointly sufficient for assault failure:39 the absence of support for the assault by the Armed Forces and the absence of support for the assault by Capital. In all instances when the assault failed, this causal combination was present. In all instances when this causal combination was present, assaults failed. This, again, signifies three things: (1) if the Armed Forces and Capital both refrain from supporting the president, the assault will fail; (2) the only way to defeat an executive assault in Ecuador is for both the Armed Forces and Capital to refrain from supporting the president; and (3) the positions adopted by the other four actors do not have consistent effects on assault failure.40 This means that no branch of the armed forces was either actively repressing the assaulted or openly supporting the assault and that the five sectors of capital were openly divided over the assault, neutral about it, or silent. Yet this is an odd result. In effect, it says that the president loses when he has no coalitional partners, but no particular coalitions of opposition actors can defeat assaults. I interpret this finding at length below. Table 8: Necessary and Sufficient Causes of Assault Outcomes Sole necessary and sufficient cause of assault VXFFHVV: Armed Forces Support Two necessary causes jointly sufficient for assault IDLOXUH: Absence of Support by Armed Forces * Absence of Support by Capital Raggin 2001). [An underlying probability of 65% will allow me to make the claim that the support of the armed forces is usually necessary and sufficient for assaults to succeed.] It turns out that ten consistent outcomes out of ten cases does in fact allow me to make this claim with 95% confidence ( =.05) 39 Goertz & Mahoney (2004) use this language to describe this causal structure. 40 Yet now we are only examining eight cases of failed assaults. How confident can we be in these results? I again use a binomial probability test to ask (again paraphrasing Ragin) ‘What is the probability of observing eight successes out of eight trials, if the underlying probability of success is 65%?’ Again, it turns out that eight consistent outcomes out of eight cases also allows me to claim that this causal combination is usually necessary and sufficient (p=.65) with 95% confidence ( =.05). 17 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt For now, however, note that I have used these two findings to gesture at a disappointing third: the positions taken by the Law, Human Rights Networks, the Media, and the Property-less do not seem to matter for whether an assault succeeds or fails. In fact, similar – sometimes identical – opposition coalitions were associated both with successful assaults (when the armed forces supported the assault) and with failed assaults (when neither the armed forces nor capital supported the assault). Are we thus to conclude that the only relevant actors for determining assault outcomes are the Armed Forces and Capital? From the perspective of proximate causes, the answer is a resounding yes. In this sense, the support or lack of support for the assault by the Armed Forces and Capital is decisive for the outcome: they are Ecuador’s ‘democracy brokers.’ Indeed, it seems that O’Donnell & Schmitter’s tentative conclusion (1986) that transitions to democracy are held hostage to the interests of the armed forces and capital is still true in Ecuador a full quarter-century after the ‘transition’ was completed. The ability to practice democracy – to express one’s political opinion – in Ecuador continues to depend on whether these two actors choose not to oppose it.41 Discouragingly, the armed forces regularly overrule the decisions of other would-be arbiters, from new social movements to agents of horizontal accountability to human rights watchdogs. This finding should be especially sobering to those still enmeshed in debates over democratic ‘consolidation’ in the region. If we were to accept Przeworski’s (1986, 1991) understanding of democracy as institutionalized uncertainty of outcomes for all major actors, we might have to conclude that Ecuador is simply not democratic, at least when it comes to executive assaults. The substantive interests of the armed forces and capital are never really part of the democratic game. It is important to realize that the logical reduction of causal complexity entailed in this analysis includes a number of simplifying assumptions. Most obviously, it assumes that the results hold for all cases of successful assaults and all cases of failed assaults in Ecuador. Yet the causal combinations analyzed here (Table 7) do not represent all hypothetically possible combinations of positions that could be taken by the six actors. For example, I do not observe any cases in which the Armed Forces support the assault; Capital does not support the assault; the Property-less do not oppose the assault; and the Law, Human Rights Networks, and the Media do all oppose the assault. Yet my results imply that this combination (were it ever to occur) would lead to a successful assault, as it includes the support of the Armed Forces for the assault. Since I have no actual cases of this 41 In this sense, they act as veto players (see Tsebelis on this usage) 18 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt combination, however, I am unable to empirically confirm that implication. I can only suggest it is reasonable given the information I have about the eighteen assaults that did occur. There is a second, potentially serious limitation of the data analyzed here: I observe no cases of assault where Capital supports the assault and the Armed Forces do not support the assault. More specifically, I do not observe any failed assaults when Capital supported the assault, but the Armed Forces did not. Had I observed this combination, the process of logical reduction employed above would have shown the position of Capital to be just as irrelevant to failed assaults as are the positions of the Law, the Media, Human Rights Networks and the Property-less. If this were the case, I could present a nice symmetrical argument where all that matters for assault outcomes is the position of the Armed Forces. (That is, when the Armed Forces support an assault, it succeeds. When they do not support an assault, it fails.) Instead, I am left with logically troubling results in which the cause of assault failure is not the simple inverse of the cause of assault success: If the support of the Armed Forces is all it takes for an assault to succeed, why isn’t the absence of Armed Forces support all that is needed to thwart an assault? The short and easy answer would be, again, that I just did not observe enough assaults. Given a larger sample of assaults, I might expect to observe the missing combination that would allow me to rule out my finding that the absence of support by Capital is a necessary condition of assault failure. But just because an answer is short and easy, does not mean that it is right. A closer look at the cases provides good reason to believe that the Capital result is an important empirical finding, rather than the artifact of insufficient observations. IV. The Dynamics of Anti-Assault Conflicts Thus far the substantive results I have uncovered paint a grim picture of Ecuadorian democracy. Yet they beg the question: ‘if all that matters is the positions taken by the Armed Forces and Capital, what determines those positions?’ answered. 1. Why are the Armed Forces supporting the president during successful assaults? 2. Why are the Armed Forces not supporting the president during failed assaults? 3. Why is Capital not supporting the president during failed assaults? Is this a spurious finding, the result of insufficient data? (That is, are the Armed Forces the only group that matters for assault outcomes?) 4. If the Capital finding is not spurious, why do the lack of military support and the lack of capital support always occur together during failed assaults? More specifically, four questions need to be 19 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt To address these questions we must reconsider the parameters of the analysis. Thus far I have used snapshots of the final positions adopted by actors to explain assault outcomes: any and all sense of process has been systematically eliminated from the analysis. In order to generate a more robust explanation of assault outcomes, I now relax that assumption and re-introduce process into this story. Sometimes the position of the Armed Forces and Capital towards an assault is clear a priori and remains constant throughout the conflict.42 In these cases, the outcome of the assault is decided deus ex machina: external conditions pre-determined the outcome by provoking strong, early reactions in Capital and the Armed Forces. In such moments, no one can do anything to change the basically pre-determined outcome of the assault; agency is a farce. Yet this happens infrequently in Ecuador. These actors’ positions are in fact shaped over the course of the conflict as the associations and individuals that make up society interact with each other. As preliminary evidence of this, consider the timing and substance of the declarations and actions of Capital and the Armed Forces for each of the eighteen executive assaults I observe in Ecuador. The Armed Forces’ position towards the assault was determined exogenously in only two of the twelve cases when it emerged as an actor. And Capital’s position was determined exogenously in none of the six cases when it took a position. In other words, the positions of the Armed Forces and Capital, as well as other grand actors, usually develop over the course of the executive assault. That is, they are endogenous to the conflict itself. I take this as prima facie evidence for the plausibility of a claim that different anti-assault tactics may have different effects on the positions taken by Armed Forces and Capital. If their positions develop during the assault, it seems likely that some responses to the assault are more likely to provoke their wrath than others. What you choose to do to defend yourself against an assault may, in other words, be the crucial determinant of whether or not you are able to regain your political voice, over the president’s objections. Here I offer a possible explanation based on casestudies of the eighteen executive assaults I observe in Ecuador. Street Protest All successful assaults include the support of the armed forces, begging the question, why are the organizations that make up the Armed Forces so willing to support the assault? It turns out 42 For such reasons as their historical relationship with the president, their historical relationship with the assaulted individuals, or the specific aspects of the context in which the assault takes place like crisis, and internal divisions/battles among sectors. 20 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt that in 80% of successful assaults (8/10) in Ecuador, the assault was primarily contested using massive street protests. Yet a mere 25% of failed assaults (2/8) were contested with street protest. Phrased differently, mass protest was associated with failed assaults 80% of the time. What might be going on here? In short, massive street protest creates public disorder, almost by definition. Traffic is blocked, everyday life is disrupted, the power of the state and its commander-in-chief are challenged. This type of activity tends to antagonize Armed Forces, constitutionally charged with the mission of and trained in the maintenance of public order. Even the Ecuadorian Armed Forces, which currently and historically have harbored deep sympathies for the plight of the protesting property-less43 seem to perceive massive protest as disruptions of public order – a threat that falls within their jurisdiction. As such, they tend to comply when the president orders them to break up protests. Thus I uncomfortably would suggest that street protest is usually an ineffective means of defending oneself against an executive assault.44 Whatever other goals it may have – and there are many – street protest is usually useless for defeating assaults. Capitalist-Dependent Tactics All failed assaults are associated with the joint absence of military support and capital support. Again the cases provide a plausible interpretation of this finding: Though Capital (understood as significant portions of Ecuadorian capitalist associations) never took a position against an assault, a full 75% of failed assaults (6/8) were contested by the opposition using the resources of some particular corporation or capitalist association.45 The support of particular capitalists for the opposition seems to take two basic forms: media tactics and legal tactics. Media tactics entailed vocal and well-financed opposition by the capitalists behind major print and television outlets, often through media owners’ organizations. Legal-institutional tactics involved funding expensive legal defenses of the assaulted in state legal institutions.46 Comparing across outcomes, assaults failed in 43 In Ecuador, it is usually the Property-less who take up street protest. Given that those protests are often over regressive economic policy, significant parts of capital often side with president, too. This might lead to the alternative hypothesis that it is not the act of mass protest that leads to the assault success, but rather (1) the fact that it is the Property-less who are protesting or (2) the fact that they are usually protesting orthodox economic policy. Yet I do not believe either of these is the case. First of all, repression of the Property-less by the Armed Forces is less likely to occur in Ecuador than probably any other South American country except Venezuela, given the nature of the Ecuadorian Armed Forces today. Moreover, when capital chooses to adopt protest as its primary means of contesting assaults, it also tends to lose (there is especially strong evidence for this in recent events in Venezuela). 44 Nonetheless, protest does not always fail. The interesting question is thus why protest occasionally works, rather than the fairly obvious reasons why it usually fails [I explore this in Chapter 4 of the dissertation]. 45 Conversely, a mere 10% (1/10) of successful assaults were contested with the resources of some capitalist organization. 46 In Venezuela, some capitalists even funded in a parallel electoral court in an attempt to keep the state electoral commission from perpetrating fraud. 21 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt six of the seven instances when these types of organizations supported the opposition. In Ecuador, it seems, overcoming assaults seems to require the participation of capitalists. Why does neither of the capitalist-dependent tactics provoke the Armed Forces into supporting the assault. Here I would return to the Armed Forces’ commitment to maintain public order. Unlike mass protest, media and legal-institutional tactics are difficult to frame as clear threats to public order. As such, the Armed Forces usually lack a convincing rationale to get involved in the conflict.47 This provides a substantive interpretation of what appeared at first glance a logical fallacy: Ecuadorian assaults fail because some specific capitalists mobilize their resources against the president using tactics that do not antagonize the Armed Forces. In such situations, I observe the societal actor “Capital” as divided or uninvolved during each failed assault, and the Armed Forces as uninvolved or neutral. Table 9: An Endogenous Theory of Assault Outcomes in Ecuador? Tactics Adopted (Endogenous to Assault Conflict) Proximate Causes Reaction of Democracy Brokers Armed Forces Support Assault (Mechanism: Perceived Threat to Public Order) Neither Capital nor Armed Forces Support Assault (Mechanism: No Perceived Threat to Public Order & Some Capitalists Assist in Anti-Assault Effort) Outcome Assault Succeeds Assault Fails Street Protest Capitalist-Dependent * Media-Based * Legal-Institutional V. Coalitional Approaches and Democracy in South America A substantial part of this paper was devoted to developing a coalitional approach to assault conflicts. The study of assault conflicts in Ecuador involves the comparison of approximately twenty cases of public conflict between scores of different associations over a twenty-five year period. Here I sought to develop a theoretical and methodological framework that enabled me to compare these cases without imposing assumptions that unjustifiably warped reality. Whether or Cameron (1998) has suggested that the perception of threat by the Armed Forces is key to understanding macro-level breakdowns of democracy. He argues that the difference between successful and failed autogolpes is the extent to which the armed forces perceived a security threat at the moment when the president attempted the autgolpe. Where the threat existed (Peru), the autogolpe succeeded. Where it did not (Guatemala), the autogolpe failed. I believe my findings partially confirm the Cameron hypothesis at the micro-level, though it is more complicated than he suggests (see also Alexander 2002). 47 22 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt not this paper has succeeded depends in large part on the extent to which my audience is convinced by the reasonableness of this framework Using this framework, I gestured at more substantive findings about assault outcomes. When do assaults succeed? When do assaults fail? I demonstrated that the outcome of an assault in Ecuador ultimately depends on the positions taken by the Armed Forces and Capital. In the end, the positions of the Property-less, the Law, the Media, and Human Rights Networks simply do not have reliable or consistent effects on assault outcomes. I also suggested that the positions taken by the Armed Forces and Capital is directly related to the tactics adopted by the assaulted individuals to contest the assault. Street protest nearly always dooms the assaulted by provoking the support of the Armed Forces for the president. Defeating assaults paradoxically requires the participation of capitalist associations in the defense of the assaulted – yet Ecuadorian Capital has never adopted a united position against an assault. It is an understatement to say that I am discouraged by these findings. Indeed, the analysis raises unsettling questions about the nature of democracy in Ecuador and other South American countries today. Can a series of assaults cripple the ability of certain groups to get involved, making future assaults unnecessary? If so, is the regime undemocratic? Has this happened in Ecuador or elsewhere in the region? Are certain groups being systematically shut out of politics each time they try to express themselves? If so, is the regime undemocratic? Has this happened in Ecuador? Are substantive interests of the armed forces and capital ever really in play? If not, is the regime undemocratic? Has this happened in Ecuador? I lack space to take up any of these questions at length here. Yet I would briefly argue that Ecuador and many of its South American neighbors might be best understood as what we have long understood and analyzed productively as “limited democracies.”48 This is to say there are certain limits on democracy’s boundaries. Previous versions of limited democracy put formal limits on participation, most famously gender-, property-, race-, and literacy- based restrictions of suffrage. Unlike these forerunners, the new limits of democracy are more fluid, yet always hover in the background. My findings suggest that these limits are defined largely by (1) the willingness of capitalists to invest in the protection of political liberty on an ad hoc basis and (2) the willingness of the general public to refrain from mass protest.49 48 On limited democracies see, e.g., Linz 2000; R.Collier 1999. I recognize that the notion of “democracy with adjectives” has been a bit of a poison pill since Collier & Levitsky’s (1997) critique, but it seems quite appropriate here. 49 Adam Przeworski has been making parallel arguments for quite some time (see, e.g., 1986, 1991). 23 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Is this situation sustainable? Recent trends suggest not. Sending presidents into exile before their terms are up has become the norm in the region: Bucaram and Mahuad in Ecuador,50 Sánchez de Lozada in Bolivia,51 and (ironically) Peru’s Alberto Fujimori. Repeated attempts to depose Venezuela’s Chávez have failed.52 Contrary to en vogue economic interpretations, I would argue that this wave of presidential deposals has been made possible by two decades of lackluster and patchy support by South American presidents for democratic freedoms – observed here as executive assaults. Given that presidents have been willing to turn a blind eye to these basic rules of the democratic game when they inconvenience executive policy, citizens are feeling increasingly less obligation to respect the outcome of elections when executive policy inconveniences them. It is all too possible that this trend could snowball into a total collapse of respect for basic rules of the democratic game. If South American democracies meet their demise in the near future, I suspect it will not be via transformations into “hybrid authoritarianism” but rather via descents into violent anarchy. Hobbes’ state of nature – rather than his tyrannical sovereign – presents the greatest threat to democracy on the continent today. Ironically, it is the actions of would-be tyrants that will have brought on this anarchy. Also Borja’s (Princeton-educated) Vice President, Alberto Dahik in 1994. Earlier this year (2004) rumors of mass protests to demand President’s Gutierrez’s resignation floated through the press. 51 Rumors surrounding a plot to depose Bolivia’s current president, Carlos Mesa, led to cabinet shuffling earlier this year (2004). 52 In addition, President Carlos Andres Perez was impeached in 1993 (as was Brazilian president Fernando Collor de Mello). 50 24 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt APPENDIX: Coding Guidelines for Reducing Ecuadorian Associations into Six Actors Each assault was analyzed in the following manner. I began by reading journalistic accounts of the assault in El Comercio (Quito), El Universo (Guayaquil), El Hoy (Quito), &/or El Telegrafo (Guayaquil). By tracing the process by which the assault unfolded, I developed a more complete understanding of the assaults I had identified in the dataset. Moreover, this allowed me to identify the public stances taken by different state and social actors on the executive assault and the issue to which it related: First, representatives of different associations interested in the assault outcome are regularly quoted in periodicals as the assault unfolded. Second, Ecuadorian associations traditionally publish public statements of their political positions in the major newspapers. Using these two types of information, I was able to identify which organizations were publicly pressuring the president and the positions those organizations adopted. When necessary and possible, I then conducted interviews with representatives of these associations to clarify the positions they took during the conflict. For each case, I thus developed a list of which associations made public statements about the assault, the position they took, and the intensity of that position. I then classified each organization as belonging to one of eighteen groups: commercial capital, agricultural capital, industrial capital, financial capital, media capital, military, police, academia/students, church, international NGOs, domestic NGOs, urban working class, indigenous, transportistas, journalists, media NGOs, state legal institutions, and lawyers. Next, I further reduced these eighteen groups into the six actors used in the above analysis. This process is depicted in Figure 1 (below). Table 10 below identifies the associations included in each group, and the aggregation of groups into the six actors. Table 11 below sketches the coding rules used to decide whether an actor took a position during each assault. The positions taken by a group were determined using the case-specific information I had regarding each assault conflict. Figure 1: Simplification of Ecuadorian Society (Org = Organization) ORG. ORG. ORG. ORG. ORG. ORG. Group (Lawyers Orgs) Actor (Law) Group (State Legal Orgs) Table 10: Composition of Societal Actors in Ecuador (continued next page) Propertyless Capital Urban Working Class Commercial UNE Federacion Nacional de Camara de Comercio de CEDOC Ecuador CTE Federacion de Exportadores de Ecuador CEOSL Camara de Construccion de Guayaquil Federacion de Trabajadores de Pinch Camara de Comercio de Quito FUT Federacion Nacional de Camara de Comercio de Federacion de Tripulantes de Ec Ecuador Oil workers 25 DRAFT: October 19, 2004 Please do not circulate without author permission. Agricultural Camara Nacional de Agricultura Camara de Agricultura de Zona I Industrial Camara de Industira Pequena de Guyaquil Camara de Industira de Guyaquil Camara de Comercio de Guyaquil Camara de Industria de Pinchichinca Camara de Pequena Industria de Pinchincha Financial Asociación Bancaria IMF WB BID Media AEDEP Television owners Informal Newspaper owners coalition pre 1985 Defeating Executive Assaults in South America William T. Barndt Health workers (FUTS) Electric workers CMS Transportistas Federacion Nacional de Choferes Federacion Nacional de Cooperativas de Transporte Federacion Nacional de Pasajeros de Ecuador FENACOTIP Federacion Nacional de Transportistas de Ecuador Federacion de Transporte de Ecuador Federacion de Transportistas Urbanas de Guyas Federacion de Cooperativas y Companias de Transporte Livino Selecto Transportes Urbanas de Pichincha Indigenous Fenoc Ecuarunari CONAIE FENACLE Fenuassc Pachakutik CMS Media Journalists Colegio de Periodistas de Pichincha FENAPE UNP Media Capital Television Owners AEDEP Informal Newspaper owners coalition pre 1985 Media NGOs SIP AIR Law State Institutions CSJ Congress (only counted when votes taken) TGC TSE Defensoria del Pueblo Alcaldesías (amparo powers) Lawyers Colegio de Abogados de Guayaquil Colegio de Abogados de Quito Federacion Nacional de Abogados Armed Forces Branches of Military National Police Special Federal Police Force Transport Police City Police Squads (Quito, Guayaquil, Cuenca) Human Rights Networks Academia/Students FEUE FEJE FESE Profesores de la Univ Central Head of Salesian Univ Church Archbishop de Quito Archbishop de Guayaquil International NGOs Election Observation Teams Amnesty International Human Rights Watch Domestic NGOs 26 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Table 11: Condensed Coding Rules Capital Supports Armed Forces Support Property-less Opposes Law Opposes Human Rights Networks Oppose Media Opposes 1. None of the five groups (industry, commerce, agriculture, finance, information) oppose the assault 2. At least two groups support the assault. 1. At least one branch of armed forces supports assault 2. There is no open or veiled dissent by any other branch 1. Two of three property-less groups oppose the assault (indigenous, urban working class, transport) 2. No group supports assault 1. No institution (or lawyers) supports or is divided over the outcome 2. At least one institution opposes the outcome 1. Open opposition by two of four groups (Academia/Students, Church, International NGOs, Domestic NGOs) 2. No group supports the assault. 1. Two of three groups (owners, journalists, NGOs) oppose the assault 2. No group opposes the assault. 27 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Works Cited Adcock, Robert and David Collier. “Measurement Validity: A Shared Standard for Qualitative and Quantitative Research. American Political Science Review 95: 3 (September 2001): 529-546. Alexander, Gerard. The Sources of Democratic Consolidation. Ithaca: Cornell University Press, 2002. Ames, Barry. Political Survival: Politicians and Public Policy in Latin America. Berkeley: University of California Press, 1987. Bates, Robert H. Markets and States in Tropical Africa: The Political Basis of Agricultural Policies. Berkeley: University of California Press, 1981. Berg-Schlosser, Dirk and Jeremy Mitchell, eds. Conditions of Democracy in Europe, 1919-39: Systematic Case Studies. London: MacMillan Press, 2000. Bermeo, Nancy. Ordinary People in Extraordinary Times: The Citizenry and the Breakdown of Democracy. Princeton, NJ: Princeton University Press, 2003. Bollen, Kenneth A. & Pamela Paxton. “Subjective Measures of Liberal Democracy.” Comparative Political Studies 33:1 (February 2000): 58-86. Braumoeller, Bear and Gary Goertz. “The methodology of necessary conditions.” American Journal of Political Science 44 (2000): 844-58. Cameron, Maxwell A. “Latin American Autogolpes: dangerous undertows in the third wave of democratization.” Third World Quarterly 19:2 (June 1998): 219-39. Cameron, Maxwell A. “Political and Economic Origins of Regime Change in Peru: The Eighteenth Brumaire of Alberto Fujimori” in The Peruvian Labyrinth: Polity, Economy, Society, Eds. Maxwell A. Cameron and Philip Mauceri, University Park, PA: Penn State Press, 1997: 36-69 Carey, John M. and Matthew Soberg Shubert, eds. Executive Decree Authority. New York: Cambridge University Press, 1998. Carothers, Thomas. “The End of The Transitions Paradigm.” Journal of Democracy 13:1 (Jan 2002): 5-21. Cheibub, José Antonio. “Minority Governments, Deadlock Situations, and The Survival of Presidential Democracies.” Comparative Political Studies 35:33 (April 2002): 284-312. Cheibub, José Antonio and Fernando Limongi. “Democratic Institutions and Regime Survival: Parliamentary and Presidential Democracies Reconsidered.” Annual Review of Political Science, vol. 5 (2002): 151-79. Collier, David and Robert Adcock, “Democracy and Dichotomies: A Pragmatic Approach to Choices About Concepts.” Annual Review of Political Science 2:1 (1999): 537-66. Collier, David and Steven Levitsky. “Democracy with Adjectives: Conceptual Innovation in Comparative Research.” World Politics 49:3 (1997): 430-51. Collier, Ruth Berins. Paths Toward Democracy: The Working Class and Elites in Western Europe and South America. New York: Cambridge University Press, 1999. Conaghan, Catherine M. and James M. Malloy. Unsettling Statecraft: Democracy and Neoliberalism in the Central Andes. Pittsburgh: University of Pittsburgh Press, 1994. Catherine Conaghan: “Politicians Against Parties: Discord an Disconnection in Ecuador’s Party System” in Building Democratic Institutions: Party Systems in Latin America, Scott Mainwaring and Timothy R. Scully, eds. (Stanford: Stanford University Press, 1995): 435-6. Dahl, Robert A. Polyarchy: Participation and Opposition. New Haven: Yale University Press, 1971. de Tocqueville, Alexis. Democracy in America [1835]. New York: Bantam, 2000. Diamond, Larry. “Thinking about hybrid regimes.” Journal of Democracy 13:2 (April 2002): 2135. Elkins, Zachary. “Gradations of Democracy? Empirical Tests of Alternative Conceptualizations.” American Journal of Political Science 44:2 (April 2000): 293. 28 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Frieden, Jeffry. Debt, Development, and Democracy: Modern Political Economy and Latin America. Princeton: Princeton University Press, 1991. Goertz, Gary and James Mahoney. “Two-Level Theories and Fuzzy Sets.” Working Paper available from Compasss/Small-N website. Version cited: January 21, 2004. Goertz, Gary and Starr (eds). Necessary Conditions: Theory, Methodology, and Applications. New York: Rowman & Littlefield, 2003. Gourevitch, Peter A. Politics in Hard Times: Comparative Responses to International Economic Crises. Ithaca: Cornell University Press, 1986. Grindle, Merilee, S. (Ed.) Politics and Policy Implementation in the Third World. Princeton: Princeton University Press, 1980. Grompone, Romeo. “Al Día Siguiente: El Fujimorismo como Proyecto Inconcluso de Transformación Política y Social” in Cotler, Julio and Romeo Grompone. El Fujimorismo: Ascenso y Caída de un Régimen Autoritario. Lima, Peru: Instituto de Estudios Peruanos, 2000. Karl, Terry Lynn. The Paradox of Plenty: Oil Booms and Petro-States. Berkeley: University of California Press, 1997. Karl, Terry Lynn and Philippe C. Schmitter. “Modes of Transition in Latin America, Southern and Eastern Europe.” International Social Science Journal 43:2 (May 1991): 269-84. Katzenstein, Peter. Corporatism and Change: Austria, Switzerland, and the Politics of Industry. Ithaca: Cornell University Press, 1984. Katzenstein, Peter. Small States in World Markets: Industrial Policy in Europe. Ithaca, Cornell University Press, 1986. Kaufman, Robert R. “Liberalization and Democratization in South America: Perspectives from the 1970s” in TFAR 1986: 85-107. King, Gary, Robert Keohane, and Sidney Verba. Designing Social Inquiry: Scientific Inference in Qualitative Research. Princeton: Princeton University Press, 1994. Latin American Weekly Report. All Issues 1978-2002. (Note that title changes slightly during the period.) Also related Regional Reports. Lawson, Chappell H. Building the Fourth Estate: Democratization and the Rise of a Free Press in Mexico. Berkeley: University of California Press, 2002. Lehoucq, Fabrice. “Electoral Fraud: Causes, Types, and Consequences.” Annual Review of Political Science 6 (June 2003): 233-56. Levitsky, Steven, “Fujimori and Post-Party Politics in Peru” in Journa of Democracy (July 1999): 82-83. Levitsky, Steven and Lucan A.Way. Competitive authoritarianism : hybrid regime change in Peru and Ukraine in comparative perspective. Glasgow, Scotland: Centre for the Study of Public Policy, University of Strathclyde, 2001. Levitsky, Steven and Lucan A. Way. “The Rise of Competitive Authoritarianism,” Journal of Democracy 13:2 (April 2002). Linz, Juan. Totalitarian and Authoritarian Regimes. Boulder, CO: Lynne Rienner, 2000/1973. Linz, Juan. “The Perils of Presidentialism.” Journal of Democracy 1:1 (1990): 51-69 Linz, Juan and Arturo Valenzuela, eds. The Failure of Presidential Democracy (Baltimore: Johns Hopkins University Press, 1994). Linz, Juan. The Breakdown of Democratic Regimes: Crisis, Breakdown, Reequilibration. Baltimore, MD: The Johns Hopkins University Press, 1978. Mainwaring, Scott & Shugart, Matthew Soberg, eds.. Presidentialism and Democracy in Latin America. New York: Cambridge University Press, 1997. Mainwaring, Scott, Daniel Brinks, and Anibal Pérez-Liñán. “Classfiying Political Regimes in Latin America, 1945-1999.” Studies in Comparative International Development 36:1 (Spring 2001): 3765. 29 DRAFT: October 19, 2004 Please do not circulate without author permission. Defeating Executive Assaults in South America William T. Barndt Marx, Karl. “The Eighteenth Brumaire of Louis Bonaparte” [1851/1852] in The Marx-Engels Reader (2nd Ed): 594-617. Robert C. Tucker, ed. New York: W.W. Norton, 1978. Marx, Karl. “The German Ideology” [1846/1932] in The Marx-Engels Reader: 594-617. Moore, Barrington. The Social Origins of Dictatorship and Democracy. Boston: Beacon Press, 1966. Munck, Gerardo L. “The regime question: Theory building in democracy studies” World Politics 54: 1 (Oct 2001). Munck, Gerardo L. and Jay Verkuilen. “Conceptualizing and Measuring Democracy: Alternative Indices.” Comparative Political Studies 35:1 (February 2002): 5-34. O’Donnell, Guillermo. “Democracy, Law, and Comparative Politics.” Studies in Comparative International Development 36:1 (Spring 2001): 7-36. O’Donnell, Guillermo. “Delegative Democracy.” Journal of Democracy 5:1 (January 1994): 55-69. O’Donnell, Guillermo. “On the State, Democratization and Some Conceptual Problems: A Latin American View with Glances at Some Postcommunist Countries.” World Development 21:8 (August 1993): 1355-69. O’Donnell, Guillermo. Bureaucratic Authoritarianism: Argentia, 1966-1973, in Comparative Perspective. Berkeley: University of California Press, 1988. O’Donnell, Guillermo, and Philippe C. Schmitter. Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies. Baltimore: Johns Hopkins University Press, 1986. Payne, Leigh A. Brazilian Industrialists and Democratic Change. Baltimore: Johns Hopkins University Press, 1994. Pérez-Líñan, Aníbal S. “Pugna de poderes y crisis de gobernabilidad: Hacia un nuevo presidencialismo?” Latin American Research Review 38:3 (2003): 149-64. Przeworski, Adam. “Some Problems in the Study of the Transition to Democracy” in TFAR 1986: 47-63. Przeworski, Adam. Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. New York: Cambridge University Press, 1991. Przeworski, Adam, Michael A. Alvarez, José Antonio Cheibub, and Fernando Limongi. Democracy and Development: Political Institutions and Well-Being in the World, 1950-1990. New York: Cambridge University Press, 2000. Ragin, Charles C. The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies. Berkeley: University of California Press, 1987. Ragin, Charles C. Fuzzy-Set Social Science. Chicago: University of Chicago Press, 2000. Ragin, Charles C. and Helen M. Giesel. “User’s Guide: Fuzzy-Set / Qualitative Comparative Analysis.” Downloaded from Charles Ragin’s homepage (University of Arizona). Version cited: October 2002. Rueschemeyer, Dietrich, Evelyene Huber Stephens, and John D. Stephens. Capitalist Development and Democracy. Chicago: University of Chicago Press, 1992. Sartori, Giovanni, ed. Social Science Concepts: A Systematic Analysis. Beverly Hills, CA: SAGE Publications, 1984. Sartori, Giovanni. The Theory of Democracy Revisited. Chatham, NJ: Chatham House Publishers, 1987. Shugart, Matthew Soberg and John M. Carey. Presidents and assemblies : constitutional design and electoral dynamics. New York: Cambridge University Press, 1992. Wendt, Alexander. Social Theory of International Politics. London: Cambridge University Press, 1999. 30