Semi-presidentialism and Democratic Breakdown

Transcription

1 Semi-presidentialism and Democratic Breakdown Jung-hsiang, Tsai Assistant Professor, Department of Political Science, National Chung Cheng University, Chia-yi, Taiwan; Paper prepared for International Political Science Association, World Congress in Santiago, Chile, July 12-6, I am very grateful to Professor Robert Elgie at Dublin City University, Ireland, Professor William Crowther at University of North Carolina at Greensboro, USA, and Professor Yu-shan, Wu, Distinguished Research Fellow and Director at Institute of Political Science at Academia Sinica (IPSAS), Taiwan, for their thoughtful comments on versions of this manuscript.

2 Abstract Why some semi-presidential democracies are more fragile leading to breakdown than other semi-presidential democracies is the major research question in this paper. Previous research asserts that strong president, cohabitation, or divided minority government contributes to democratic breakdown of semi-presidential countries. This paper, instead, proposes that the institutional factor (balance of power between the president and parliament) and the behavioral factor (coordination problems) are more crucial to account for democratic breakdown of semi-presidential countries. Consequently, this paper has explored two different causal mechanisms that foster democratic breakdown in semi-presidential countries. First, balanced semi-presidentialism which associates equal balances of power between the president and parliament with coordination problems has caused democratic breakdown of frequent occurrence. Second, powerful presidents gaining high legitimacy in semi-presidentialism are more prone to overstep their constitutional authority, thereby sliding into electoral authoritarianism. Keywords: semi-presidentialism; balance of power, coordination problems; democratic breakdown.

3 Introduction The charms of semi-presidentialism for scholars of comparative constitutional engineering are that it combines the virtues of parliamentarism and presidentialism, ameliorating the vices of them (Sartori, 1997; Pasquino, 1997, 2007). 1 However, in some countries the operation of semi-presidentialism has presented a different scenario. Presidential usurpation, parliament s intransigence, and serious deadlock between the president and parliament have been the glaring features of some semi-presidential countries. Moreover, some of the semi-presidential countries have come to the dénouement of democratic breakdown. Despite the gloomy result of political performance regarding semi-presidentialism, evidence shows that more countries have chosen semi-presidentialism as their constitutional type. Among 150 free and partly free countries, 46 countries, almost equal to one-third, have embraced semi-presidentialism (Elgie, 2007a; Freedom House 2008 Survey). To date, the number of semi-presidential countries has increased to keep pace with the popularity of parliamentarism and presidentialism. It is timely and worthwhile to examine the possible impacts of combining semi-presidentialism and particular institutional or behavioral factors on nascent democracies. This paper aims to answer why some semi-presidential countries have gone down the path of democratic breakdown, and what endogenous factors or causal mechanisms are necessary or sufficient for causing that result. 2 The Weimar Republic was the first tragic case of semi-presidential democracy. The democratic regime collapsed and was succeeded by an authoritarian regime. Political features such as double dependence (the government has to get support from the president and parliament) and divided minority government, were the major causes which triggered the breakdown of democracy in the Weimar Republic (Bernhard, 2005; Skach, 2005). Afterwards, 17 semi-presidential countries followed suit, breaking their democracy down dramatically (Kirschke, 2007, p. 1387; Moestrup, 2007, p. 42). 3 This begs a couple of questions: why some semi-presidential 1 The proponents of semi-presidentialism have mostly cited France and Portugal as the successful cases to argue that semi-presidentialism can facilitate democratic stability. 2 Elgie (2007b, p.62) argues that semi-presidentialism as a general category should not be used as an explanatory variable we need to examine the impact of varieties of semi-presidentialism. Here we follow this research guideline countries of democratic breakdown are Armenia, Belarus, Burundi, Central African Republic, Comoros, Congo Brazzaville, Germany, Guine-Bissau, Haiti, Kyrgyzstan, Niger, Peru, and Russia. Please see Moestrup (2007, p.42). Democratic Republic of Congo, Mali, Rwanda, and Togo as semi-presidentialism experiencing democratic breakdown are added to

4 democracies are more prone to breakdown than others; whether the same problems of the Weimar Republic have plagued subsequent semi-presidential democracies; which combination of institutional factors and behavioral factors is necessary or sufficient to explicate the outcome of democratic breakdown. The explanation of why semi-presidential democracies considered breakdown has to take account of how they interacted with political contexts in which they operated. Strong presidents, cohabitation, and divided minority government are three major causes contributing to the breakdown of semi-presidential democracies in the scholarly previous accounts. First, strong presidents, especially wielding decree powers, are more prone to bypass parliament and unilaterally enact their preferred policies. This is so called superpresidentialism which has caused democratic erosion in Russia (Fish, 1997, 2000) or in Peru (Mauceri, 2006). Second, the intraexecutive conflicts between presidents and prime ministers during periods of cohabitation have brought about military coups, ending up with democratic breakdown in 12 African semi-presidential countries (Kirschke, 2007). 4 Kirschke s study provides empirical support for Linz s thesis (1994) that divided executives can furnish the intervention of military into politics of semi-presidentialism. Third, political dynamics in the phase of divided minority government are so highly unstable that strong presidents take advantage of this situation to abuse their authority to be constitutional dictatorship in semi-presidentialism (Skach 2005). In the same token, Moestrup (2008, p. 23) proposes that in three African semi-presidential countries: Burkina Faso, Central African Republic, and Guinea-Bissau in 2003, presidents with significant powers over cabinet formation and dismissal have opted for governing through minority governments, thereby contributing to polarizing politics and setting the stage for an authoritarian reversal. However, there are two caveats for this thesis. First, minority government only can be considered as a sufficient cause or a conditional factor for democratic breakdown of semi-presidentialism. As noted elsewhere, some president-parliamentary systems such as Taiwan have the list, according to Kirschke (2007, p.1387). In addition, Moestrup (2007) concludes that nascent semi-presidential democracies are not more prone to breakdown than nascent presidential or parliamentary democracies. Moestrup focuses on comparing the survival rate of semi-presidential countries with that of presidential countries or that of parliamentary countries. This paper instead centers on what endogenous causes are prone to bring about the demise of semi-presidential democracies. 4 However, in a larger dataset, Moestrup (2007) finds that there is no statistically significant on the rate of democratic survival between the semi-presidential democracies with the experience of cohabitation and the ones without the experience of cohabitation.

5 passed the acid test of democratic survival in the period of minority governments (Wu and Tsai, 2008). 5 Second, polarizing politics or grandstanding between the president and parliament makes the governing of minority government more vulnerable for democratic breakdown, rather than the reverse as minority government causes polarization. Cohabitation and divided minority government are the factors of political dynamics. Both of them derive from the results of presidential and parliamentary elections. When opposing forces control different executive posts (president and prime minister) and struggle fiercely, this chaotic situation easily entices the military to intervene in politics and eventually causes democratic breakdown (Elgie and McMenamin, 2008; Kirschke, 2007; Linz, 1994). However, France has experienced three periods of cohabitation, but did not derail the path of democratic development. The partial reasons for that result are that the president in France, relatively speaking, has moderate powers and France has a two-alliance party system. At issue here is why cohabitation has produced various outcomes in different semi-presidential countries. What other endogenous factors can account on democratic breakdown is our concern in this paper. In addition, Elgie (2007b, p.69) proposes that divided minority government has caused 4 semi-presidential countries such as Armenia, Belarus, Burkina Faso, and Weimar Germany to democratic collapse, but it is not the major driving force for most of the failed semi-presidential democracies. Divided minority government is highly related with the party system fragmentation a semi-presidential country has. In short, if we can gauge the impacts of institutional and behavioral variations on the breakdown of semi-presidential democracies, we can offer some useful suggestions for the countries aspiring to adopt semi-presidentialism in the future. Different subtypes of semi-presidentialism have great repercussions on the rate of democratic survival. 6 According to the Shugart and Carey (1992) framework, semi-presidentialism can be categorized into president-parliamentary systems and premier-presidential systems. 7 Under premier-presidentialism, the prime minister and cabinet are exclusively accountable to the assembly majority, while under president-parliamentarism, 5 The rationale and survival of frequent minority governments in Western Europe can see Strøm (1990). 6 Regarding the subtypes of semi-presidentialism and political deadlock, related discussions can also see Tsai (2008). 7 Roper (2002) argues that some presidents in premier-presidential regimes are more powerful than others and it can cause more cabinet instability.

6 the prime minister and cabinet are dually accountable to the president and the assembly majority (Shugart, 2005, p.333). In other words, the president in president-parliamentarism is more powerful than the president in premier-presidentialism. The risk of president-parliamentarism is that it is easy to generate an ambitious president abusing power and pursuing constitutional dictatorship (Shugart, 2005; Shugart and Carey, 1992). In terms of the rate of democratic survival, president-parliamentary systems do perform worse than premier-presidential systems (Elgie, 2007b; Moestrup, 2007). As more countries have chosen semi-presidentialism as their constitutional type, it provides rich data for testing the feasibility of semi-presidentialism. In the past work on semi-presidentialism, there are three different lines of research. First, it analyzes crucial cases such as France, Portugal, and Weimar Republic in depth (Pasquino, 2007; Skach, 2005; Suleiman, 1994). Second, it constructs different subtypes to explicate stability or instability of semi-presidential democracies (Shugart and Carey, 1992; Skach, 2005; Tsai, 2008; Wu, 2000). Third, it uses some large-n datasets to detect or verify what possible factors affect the operation of semi-presidentialism (Amorim Neto and Strøm, 2006; Cheibub and Chernykh, 2008; Elgie, 2007b, 2008; Fish, 2005; Kirschke, 2007; Moestrup, 2007). Any of these approaches has its uses and limitations. This paper adopts the method of controlled comparison to deduce possible causes and causal mechanisms for democratic breakdown of semi-presidential countries (Ragin, 1987; George and Bennett, 2005; Tilly, 2001). There are two reasons why the method proposed here better captures political performance of semi-presidential countries. First, by comparing more cases, it can verify the arguments from single case studies and establish more generalizations regarding the performance of semi-presidentialism. Second, it can detect necessary or sufficient causes of the cases with the same result by comparison. In short, this paper aims to provide various causal mechanisms to explain democratic breakdown of semi-presidentialism. Additionally, in this paper, democratic breakdown refers to two different routes. One is the quick death of democracy, which means a liberal democracy transforming into an authoritarian regime by coup and the other is the slow death of democracy, which connotes a liberal democracy eroding to become an electoral democracy, and then to become an authoritarian regime (O Donnell, 1992; Schedler, 1998). Attempts such as violation of constitution norms, abuse of power, disregard of civil liberties, and excessive violence by the authorities can be

7 seen as the possible causes for democratic breakdown (Linz, 1978, p.92). The rest of this paper proceeds as follows: first, the configuration of semi-presidentialism is outlined, then the possible factors for democratic breakdown of semi-presidential democracies are analyzed, and finally, the causal mechanisms of democratic breakdown in semi-presidentialism are proposed. The Configuration of Semi-presidentialism Semi-presidentialism is defined as the situation where a country has a popularly elected fix-term president and a prime minister and cabinet who are responsible to the legislature (Elgie, 1999). Within semi-presidentialism, it can emerge different government types or phrases. Duverger (1980) describes the French regime as alternating between presidential and parliamentary phases. Following the same line of thinking, Suleiman proposes that the French constitution of 1958 was far more flexible than was initially recognized. It contains a safety valve that avoids the clash of two popularly elected legitimacies by permitting the political system to function as a presidential system, now as parliamentary (Suleiman, 1994, p.151). By contrast, Sartori objects to the alternation interpretation on the grounds that it blows apart the integrated nature of the system, and we do not have a veritable system. It is better to use the term oscillation to replace it (Sartori, 1997, p.124). No matter whether it is alternation or oscillation, the particular political system cannot be regarded as a mixed system that combines the features of presidentialism and parliamentarism. 8 There are essential differences between semi-presidentialism and parliamentarism or between semi-presidentialism and presidentialism. The distinguishing feature is the possibility of cohabitation (Bahro et al, 1998). During cohabitation, the regime is not a pure parliamentary system, since some presidents in semi-presidentialism have certain reserved powers such as veto power, dissolution power, and decree power. In a nutshell, semi-presidentialism can be described as a contingent political system, depending on the interactions between constitutional contexts and political dynamics. Hence, we may treat it as a new category to investigate its variations and political performances. 8 Many scholars, such as Przeworski, Alvarez, Cheibub, and Limongi (1996) and Cheibub (2002), prefer to use the mixed system to describe semi-presidentialism. Skach (2005) proposes semi-presidentialism is not a mixed system or combines the features of presidentialism and parliamentarism and indeed is the third type of constitutional system. In this vein, we second Skach s idea.

8 Semi-presidentialism like any other constitutional type has its merits and demerits. On the bright side, first, semi-presidentialism can provide a solution to the situation of the president facing the parliament with an opposing majority by cohabitation. Cohabitation indicates that two political leaders with different party affiliations have to share political authority. This arrangement does not necessarily avoid conflict, but it can constrain political leaders to grasp authority only in their reserved domains. Second, the president in semi-presidentialism can be a counterweight to the parliament. The parliament is more concerned a local version of public good, but the president is more focused on a national version of public good. Furthermore, the direct election of a fixed-term president can provide the system with political stability and legitimacy, even if the parliament is highly fractionalized and governments are unstable (Elgie, 2007b, p.55). Third, the fixed term of president can lessen the crises of government instability or reshuffle. When the prime minister or cabinet steps down for political reasons, the fixed-term president can still be the political authority to solve government crises and prevent them from escalating into regime crises. On the down side, first, semi-presidentialism induces conflict between the dual leaders. Protsyk (2006) argues that the intra-executive conflict between the president and prime minister is more likely when the parliament exclusively controls the power to dismiss cabinet, when the president does not share ideological orientation with the cabinet, and when a president coexists with minority cabinets. In the same token, Fish (2005, p.207) argues that semi-presidentialism per se is not democracy s antagonist but can produce some pathological effects such as policy paralysis and awkward cohabitation. Second, the president and parliament can lock horn over legislation in semi-presidential countries when they hold considerable powers but oppose each other (Tsai, 2008). Third, the president with strong powers can possibly become constitutional dictatorship such as President Yeltsin in Russia (Colton and Skach, 2005). In the past work on semi-presidentialism, it mainly focuses on subtypes of semi-presidentialism and political stability. 9 This paper turns to a different angle to pay attention to political instability or democratic breakdown. It aims to grasp what possible causes can contribute to democratic breakdown in semi-presidentialism. In addition, some research has focused on the general 9 Certainly, some studies are exceptional such as Bernhard (2005) and Skach (2005). However, they only focus on some crucial cases. This paper extends to almost all the cases of democratic breakdown in semi-presidentialism.

9 pattern of political performance in semi-presidentialism, especially for nascent democracies (Elgie, 2007b; Moestrup, 2007). Following the same line, this paper attempts to analyze why some nascent semi-presidential countries suffering democratic breakdown. Many causal factors such as strong military intervention, severe economic crises, polarized political competition, and specific institutional features can bring about the breakdown of democracy. In this paper, we distinguish the factors into two categories. Military intervention and economic crises are treated as non-institutional factors, while the balance of power between the president and parliament is an institutional factor. First, we have to deal with what non-institutional factors are necessary or sufficient for democratic breakdown in semi-presidentialism. Second, we take stock of the institutional factor combined with the behavioral factor contributing to the result of democratic breakdown. The Possible Factors: the Military and Economy The military can play a vital role in the demise of a semi-presidential democracy. Linz (1994, p.57) argue that A constitutional and political problem connected with the dual executive model that deserved considerable attention is the question of who has authority over the armed forces, the president or his prime minister. The conflictual situation is like Once a personal relationship is established between the president and the minister of defense and the heads of the military establishment, the president is likely to see any interference in that relationship by the prime minister as undesirable and to jealously guard the autonomy of the military establishment (Linz, 1994, p.58). Moreover, when the dual leaders of semi-presidentialism are loggerheads with each other, it creates an incentive for the military to intervene in politics, undermining democratic practice (Elgie and McMenamin, 2008). First, we anticipate to test whether the military in semi-presidential countries choose to intervene in the situation in which the president and prime minister come to a head and in the end it causes the breakdown of a democratic regime. Kirschke s study (2007) on African semi-presidential countries finds strong support about the correlation between cohabitation and military coup. Can this proposition be also true in other semi-presidential democracies? Second, there are some other reasons than cohabitation for the military intervening in politics. Hence, we also enumerate those factors within the semi-presidential countries suffering democratic breakdown. In summary,

10 here we plan to show whether or not the frequency of military intervention in semi-presidential countries is high or low and what factors can bring about democratic breakdown. Democratic breakdown indicates military coup or the onset of authoritarianism. Table 1: Semi-presidential Countries and Democratic Breakdown Country Political situation Transition Year Armenia Strong president Electoral authoritarianism 1996 Belarus Strong president Electoral authoritarianism 1994 Burundi Cohabitation Military coup 1993, 1996 Central African R Cohabitation Military coup 2003 Comoros Cohabitation Military coup 1995, 1999 Congo Brazzaville Cohabitation Military coup 1997 Democratic R. Congo Cohabitation Military coup 1997 Guine-Bissau Cohabitation Military coup 1999, 2003 Haiti Strong president Military coup 1994 Kyrgyzstan Strong president Electoral authoritarianism 1995 Mali Cohabitation Military coup 1991 Niger Cohabitation Military coup 1995 Peru under Fujimori Deadlock Electoral authoritarianism 1992 Russia under Yeltsin Deadlock Military coup 1993 Russia under Putin Strong president Electoral authoritarianism 1999 Rwanda Cohabitation Military coup 1994 Weimar Republic Deadlock Electoral authoritarianism 1933 Source: Fish (2005); Kenny (2004); Kirschke (2007); Matsuzato (2006); Mauceri (2006); Roeder (1994); Skach (2005). According to Table 1, we can find three patterns. First, semi-president countries with a strong president end up with electoral authoritarianism such as Armenia, Belarus, Kyrgyzstan, and Russia under Putin. Second, affected by the experience of divided executives (cohabitation), military coup has been the feature of nine semi-presidential countries. Third, serious deadlock between the president and parliament is responsible for democratic breakdown in Peru under Fujimori ( ), Russia under Yeltsin ( ), and Weimar Germany. Regarding the correlation between cohabitation and military intervention is 11 out of 17 (64%). Simply put, semi-presidentialism with cohabitation is more likely to bring about military intervention, but this is not necessary for there to be the only scenario for democratic breakdown in semi-presidential countries. Another possible factor which impacts on democratic performance of semi-presidentialism is economic development. Semi-presidential countries with a lower GDP per capital figure failed in democratic performance, while semi-presidential countries with a higher GDP per capital figure mostly succeeded (Elgie, 2007b, p.70). The level of economic development has a strong impact on the performance of semi-presidential democracies. However,

11 using the military or economic variable to explain cannot tell us what institutional and behavioral factors are more prone to cause democratic breakdown. This paper takes a step further to parse out institutional factors and behavioral factors which foster democratic collapse in semi-presidential countries. 10 Causal Mechanisms of Democratic Breakdown Students of democracy have been concerned about why some democracies can be self-enforcing or not self-enforcing and how different incentive structures induce democratic stability or instability. On the one hand, when citizens can solve a coordination problem to defend the limits on the state, then the sovereign would choose to respect the constitution or rule of law (Weingast, 1997, 2004). On the other hand, when citizens have a coordination dilemma such as holding different views about limits on the state or being unwilling to defend these limits, then the sovereign can transgress the constitution without any serious challenge from the mass level (see Weingast, 1997, p.246). The establishment of any regime, no matter which is a democracy or an authoritarian regime, needs support of its citizens. Various regime preferences of the citizens profoundly affect strategic calculations and actions of political leaders to bolster democracy or erect authoritarianism (Alexander, 2002). This view argues that the support of certain citizens such as the rightist voters can assist authoritarian rule or sustain democratic consolidation in some Western European countries. Similarly, one study emphasizes that the cooperation between part of the divided elite and part of the masses can generate the opening of democracy and the conflict between the unified elite and the masses demanding for change can induce the deepening of authoritarianism (Yashar, 1997). In addition, some studies focus more on the competition of political elites. After tracing the trajectories of post-communist democratization, McFaul (2002) proposes that power struggles between split but balanced elites have lead to democratic instability or even the rise of dictatorship. Based on the experiences of multi-party countries in Latin America such as Brazil and Chile, one study suggests that polarization occurs when the moderate left closes ranks with the extreme left and the moderate right ties up with the extreme 10 Shugart (2005, p.327) argues that An institutional approach defines the authority patterns of the executive and assembly and how they are constitutionally related to one another; a behavioral approach, on the other hand, focuses on extra-constitutional factors such as the party system and leadership dynamics. Here we follow Shugart s schema to analyze the different causes for democratic breakdown in semi-presidentialism.

12 right, entailing the collapse of democracy in the two countries (Cohen, 1994). Applying these theses to semi-presidentialism, we can get a more complicated scenario since there are more than one sovereign such as the president and parliament. Within semi-presidentialism, we can identify four possible actors: the president, the parliament, the citizens supporting the president, and the citizens supporting for the parliament. How these four agents (players) interact to instigate democratic breakdown intrigues us here. This paper aims to explore why some semi-presidential democracies are not self-enforcing and how certain incentive structures can produce more democratic instability or even democratic breakdown in semi-presidential countries. Basically, this paper proposes that both the institutional factor (balance of power between the president and parliament) and the behavioral factor (coordination problem) can account for the collapse or instability of semi-presidential democracies. The Institutional Factor Relations between the president and parliament in semi-presidentialism have a strong impact on legislative output and political stability. 11 The worst scenario is an overweening president vis-à-vis an intransigent parliament. The president with unilateral legislative powers can bypass the parliament to enact his preferred policies while the parliament with a majority of opposing him can rescind his orders and pass new laws in lieu of them. When both the president and parliament flex their constitutional muscles, serious deadlock culminates and democratic breakdown entails afterwards. The best scenario is a negotiable president vis-à-vis a workable parliament. The president holding considerable powers is willing to make compromise with the parliament on 11 Cox and Morgenstern (2001) offer a rational choice perspective to analyze relations between the president and parliament in presidentialism. There are four types of the interactions: (1) imperial president versus recalcitrant assembly (when president undertakes unilateral action, the assembly rejects); (2) Coalitional president versus workable assembly (when president uses bargaining, the assembly also use bargaining); (3) Nationally-oriented president versus venal or parochial assembly (when assembly demand payments, the president pays off); (4) Dominant president versus subservient assembly (when the president dictates, the assembly is acquiesce (Cox and Morgenstern, 2001, pp ). In parliamentarism, due to the fusion of the executive and legislative branches, relations between them depend on intra-party and inter-party competition (King, 1976; Strøm 1989). In semi-presidentialism, relations between the executive and legislative branches are more complicated. When the president s party has a majority in parliament, relations between the executive and legislative branches can be more workable. When cohabitation (the president and prime minister come from opposing parties) or minority government occurs, relations between the president and parliament can be more conflicting.

13 legislation. 12 In other words, the president would make policy concessions in exchange for the parliament s support and the parliament would revise policy proposals to consider the president s preferences in order to get his signature for the promulgation of laws. Constitutional practice exhibits the relativity of power such that any branch s power cannot be enforced independently, and it has to weigh other branch s powers or political strengths by comparison in order to assess its own power. In order to understand constitutional practice in any nation, one must explore the interaction of different branches instead of focusing on any single branch. The relative strength between the executive and legislative branch translates into a measure of how many powers the president or the parliament can use to respond to each other. The relative powers of presidents and parliaments in semi-presidentialism can be traced along four dimensions. First, regarding the formation of government, in semi-presidentialism the president nominates the prime minister subject to consent of the parliament. 13 Second, on the termination of government, the parliament in semi-presidentialism can pass a vote of no-confidence to censure the cabinet. Third, on legislation, when the parliament passes the bills which the president dislikes, he can veto them and let the parliament reconsider them. Fourth, the president can enact some decree orders in lieu of laws passed by the parliament while the parliament can pass new laws by a simple majority to replace the president s decree orders. Under semi-presidentialism, interactions between the president and parliament hinge on exercising some relative powers such as cabinet formation, cabinet dismissal, dissolution of parliament, veto or override power. First, presidential powers related to interactions between the executive and legislative branches can be separated into five categories. We will give each a value of 0, 1 or 2. On cabinet formation, when the president can entirely decide the premiership, we score this a 2 for the presidential power index. When the president names the premier 14 but has to get the approval of parliament (investiture), we score a 1 for presidential power index. Where the parliament can independently decide the premiership by a vote, we score this a 0 for the presidential power index. Regarding minister dismissal, when the 12 The relationship between the president and parliament in presidentilaism is transactional. Related discussions please see Shugart (2005) and Carroll and Shugart (2007). 13 However, in some semi-presidential countries the president can nominate the premier without the consent of parliament such as Taiwan. 14 Here we use prime minister and premier interchangeably.

14 president can dismiss any cabinet minister at will, we score this a 2. When the president and the prime minister both decide to dismiss cabinet ministers, such case is scored a 1. Where the president does not have any right to dismiss cabinet ministers, we score this a 0. On dissolution of parliament, when the president can exercise this power unrestricted, we score this as a 2. When the president can exercise dissolution power restricted by frequency or point within term, such case is scored a 1. When the president does not hold dissolution of power, we score it as a 0. On decree power, when the president holds constitutional decree power and delegated decree power by parliament, we score it as 2. When the president only has constitutional decree power, we score it as 1. When the president does not hold any constitutional decree, we score it as 0. On veto power, when the president holds package veto and partial veto power, such case is scored as a 2. When the president only has partial veto power, such case is scored a 1. When the president does not hold any veto power, we score this case with a 0. On proposal of referenda and judicial review, when the president can appeal bills passed by the parliament to referenda or judicial review, such case is scored as a 1. When the president does not possess that power, we score it as a 0. Table 2 indicates the construction of the index of presidential power regarding interactions between the executive and legislative branches. Parliamentary powers related to interactions between the executive and legislative branches can be classified into five groups. Regarding cabinet formation, when parliament can independently decide the premiership, we score this as a 2 on the parliamentary power index. When the parliament holds investiture power, we score this as a 1. When the parliament does not possess investiture power, we score this as a 0. On cabinet survival, when the parliament can propose censure motions without any restriction, we give such a case a score of 2. When the parliament may censure, but it has to face the fate of dissolution, we score such case with a 1. When the parliament may not censure, we score such case as a 0. On minister dismissal, when the parliament may dismiss any individual minister, we score it as a 2. When the parliament may dismiss any individual minister, but has to propose an alternative minister, we may score it as a 1. When the parliament may not dismiss any individual minister, we score it as a 0. On decree power, when all the presidential decrees have to be reviewed by the parliament, we score this as a 2. When the presidential decrees are enforced after a certain period and then are reviewed or suspended by the parliament, we score this as a 1. When the parliament has to pass laws in lieu of presidential decrees, we score such

15 case as 0. On override power, when the parliament may override presidential vetoes by a simple majority of deputies, we score this as a 2. When the parliament may override presidential vetoes by an absolute majority of deputies, we may score this as a 1. When the parliament may not override presidential vetoes, we may score this as a 0. Table 3 shows the construction of the index of parliamentary power regarding interactions between the executive and legislative branches.

16 Table 2Presidential Power Index Cabinet formation 0 The parliament can independently decide the premiership by a vote 1 The president names the premier but has to get approval from parliament 2 The president can entirely decide the premiership Minister Dismissal 0 The president does not have any right to dismiss cabinet ministers 1 The president and the prime minister both decide to dismiss cabinet ministers 2 The president can dismiss any cabinet minister at will Dissolution of Parliament 0 The president does not have dissolution of power 1 The president can exercise dissolution power restricted by frequency or point within term 2 The president can exercise this power unrestricted Decree Power 0 The president does not hold any decree power 1 The president holds constitutional decree power 2 the president holds constitutional decree power and delegated decree power by the parliament Veto Power 0 The president does not have any veto power 1 The president holds package veto 2 The president holds package veto and partial veto power Proposal of Referenda or Judicial Review 0 The president does not have this power 1 The president can appeal bills passed by the parliament to referenda or judicial review,

17 Table 3: Parliamentary Power Index Cabinet formation 0 The parliament does not have investiture power 1 The parliament has investiture power 2 The parliament can independently decide the premiership Cabinet Survival 0 The parliament may not censure 1 The parliament may censure, but it has to face the fate of dissolution 2 The parliament can propose censure motions without any restriction Minister Dismissal 0 The parliament may not dismiss any individual minister 1 the parliament may dismiss, but has to propose alternative minister 2 The parliament may dismiss any individual minister Reviewing Decree Power 0 The parliament has to pass laws in lieu of presidential decrees 1 The decrees are reviewed by the parliament after a period of time 2 All the presidential decrees have to be reviewed by the parliament Override Power 0 The parliament does not have override power 1 The parliament may override vetoes by absolute majority of deputies 2 The parliament may override vetoes by simple majority of deputies

18 According to Elgie (2007a, 9), there are fifty-five semi-presidential countries. When semi-presidential countries are not democratic, the constitutional type means nothing but a written document. Following Freedom House s survey of 2008, we can exclude 19 semi-presidential countries which are not free on the indicators of political rights and civil liberties. 15 In addition, in order to control the effect of democratic duration, we only choose the nascent semi-presidential countries to be the scope of research. 16 By combining two indexes, Table 4 and Table 5 show the balance of power between the president and parliament in the semi-presidential countries experiencing democratic breakdown and the countries without democratic breakdown. On the whole, the semi-presidential countries experiencing democratic breakdown appear to have either equal balances of power between the president and parliament or the balance of power tilting toward the president. For the countries without democratic breakdown, the overall pattern is more or less the balance of power in favor of the parliament, except for some deviant cases such as South Korea and Ukraine. Table 4: Index of Presidential and Parliamentary Powers in Semi-presidential Countries Experiencing Democratic Breakdown Country Presidential Powers Total Parliamentary Powers Total A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 Armenia Belarus Burundi Central African R Comoros Congo Brazzaville Democratic R. Congo Guine-Bissau Haiti Kyrgyzstan Mali Niger Peru under Fujimori Russia under Yeltsin Russia under Putin Rwanda Weimar Republic A1 Cabinet Formation; A2 Minister Dismissal; A3 Dissolution of Parliament; A4 Decree Power; A5 Veto power; A6 Proposal of Referenda or Judicial Review; B1 Cabinet Formation; B2 15 The data are from Freedom House ( and were accessed on March 6, countries (Austria, Finland, France, Iceland, Ireland, and Portugal) are not included in this analysis. However, Weimar Germany is an exception here. We use it as a reference point.

19 Minister Dismissal; B3 Cabinet Survival; B4 Reviewing Decree Power; B5 Override Power Source: Blaustein and Flanz (1999); Kenney (2004); for the Weimar Republic, see Skach (2005). Table 5: Index of Presidential and Parliamentary Powers in Semi-presidential Countries Country Presidential Powers Total Parliamentary Powers Total A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 Bulgaria Croatia Gabon Georgia Lithuania Macedonia Madagascar Mozambique Nambia Poland Romania Sao Tome & Principe Senegal Slovakia Slovenia South Korea Sri Lanka Taiwan Ukraine A1 Cabinet Formation; A2 Minister Dismissal; A3 Dissolution of Parliament; A4 Decree Power; A5 Veto power; A6 Proposal of Referenda or Judicial Review; B1 Cabinet Formation; B2 Minister Dismissal; B3 Cabinet Survival; B4 Reviewing Decree Power; B5 Override Power Source: Blaustein and Flanz (1999). The Behavioral Factor (Coordination Problems) The behavioral factor here refers to the issue of democratic legitimacy. Semi-presidentialism possesses the feature of dual democratic legitimacy like presidentialism does. 17 Semi-presidentialism can exhibit a more conflictual relationship between the president and parliament, since presidents in semi-presidentialism are more powerful than presidents in presidentialism on an average. 18 That either the president or parliament enjoys more citizen supports than the other affects strategic interactions between them. When citizen supports for the president and for the parliament are almost equally balanced, then it is a scenario with coordination problems. When citizen supports for the president prevail over citizen supports for the parliament or 17 Linz (1994) argues that dual democratic legitimacy is the Achilles heel of presidentialism. 18 Some subtypes of semi-presidentialism such as premier-presidential systems have less powerful presidents than president-parliamentary systems. Related discussions can see Shugart (2005) and Shugart and Carey (1992).

20 the other way around, then it is a scenario without coordination problems. The best indicator for coordination problems is polling data about the support of two branches. Since polling data are not available for all the semi-presidential countries, here we adopt the power distribution in the legislature to be a proxy for the survey data about public support for the president and parliament. 19 When the president s party wins more than 60 percent of the vote, then the case is identified as the balance of power favoring the president. When the opposition party or parties wins more than 60 percent of the vote, then the case is identified as the balance of power tilting toward the parliament. Cases in which neither the president s party or the opposition wins a clear majority are classified as countries with equal balances of power. Democratic legitimacy can be a useful barometer of regime stability. However, democratic legitimacy is not a necessary and sufficient cause for maintaining democratic consolidation. Moreover, when a regime enjoys high legitimacy, political leaders with authoritarian tendency in the regime could overuse it to justify their unconstitutional rule. For example, President Alberto Fujimori launched the Autogolpe (self-coup) to overstep his constitutional authority even when he gained a high public approval of the citizens in Peru (Kenney, 2004). Table 6: Semi-presidential Countries and Democratic Breakdown Country Strong President Strong Parliament Coordination Problem Breakdown Armenia Belarus (president) 20 1 Burundi C. A. R Comoros Congo B D. R. Congo Guine-Bissau (president) 1 Haiti (president) 1 Kyrgyzstan (president) 1 Mali (president) 1 Niger Peru (Fujimori) (president) 1 Russia (Yeltsin) Russia (Putin) (president) 1 Rwanda Weimar Republic SourceFor Armenia, see Markarov (2006) and Rutland (1994); for Eastern European 19 McFaul (2004) uses a similar indicator to discuss democratic transition of post-communist countries. 20 It means that the support favors the president.

21 countries such Belarus, Kyrgyzstan, Russia, see Absolute Astronomy ( for African countries, see African Elections Database (africanelections.tripod.com), Absolute Astronomy, Kirschke (2007) and worldstatesman.org; for Weimar Republic, see Skach (2005). Table 7: Typology of Semi-presidential Countries with Democratic Breakdown Parliament President Coordination P. Balanced Kyrgyzstan Armenia, Burundi, Guine-Bissau CAR, D. R. Congo, Russia(Yeltsin), Weimar Germany S. President Belarus, Haiti, Niger, Rwanda Mali, Peru, Russia (Putin) S. Parliament Comoros, Congo B. Taken together, Table 6 is a truth table for semi-presidential countries experiencing democratic breakdown. 1 indicates the presence of the cause, while 0 connotes the absence of the cause. When the score of presidential powers or parliamentary powers is 6 or above, we code it as a strong president or a strong parliament. Table 7 indicates the creation of a two-dimensional space with the constitutional factor defining the vertical axis and the behavioral factor defining the horizontal axis. Each axis has three categories. The basic assumption here is that the more balanced between the president and parliament constitutionally and the more the coordination problems, the higher the possibility of democratic breakdown. All of the semi-presidential countries experiencing democratic breakdown can be located in each cell in the typology. There are three scenarios which can be presented here. First, six of the semi-presidential countries such as Armenia, Burundi, CAR, D. R. Congo, Russia (Yeltsin) and Weimar Germany, experiencing democratic breakdown manifest that they have the same scenario of equal balances of power between the president and parliament in tandem with coordination problems. The interactions between the president and parliament in these semi-presidential countries are similar to the Chicken game of rational choice approach, as shown in Figure The dashed line with two-headed arrows connotes there is colliding relationship between the two branches. The solid 21 Regarding the Chicken game, please see Snyder (1971).

22 line with a single arrow indicates there is supportive relationship. When both camps (pro-president and pro-parliament) have a misperception of being capable to hold an advantage over their opponents or an underestimate of their opponent s strength, they choose to have a confrontation with each other, thereby leading to the point of democratic breakdown. When opposing parties separately control the presidency and parliament and when the president and parliament both hold considerable powers, the situation is likely to produce political conflicts or deadlock between the executive and legislative branches (Tsai, 2008). For instance, in the Russian First Republic, the president holds veto power, but the parliament can override veto in a simple majority. As the Russian parliament controls the agenda-setting and the parliament has a legislative majority against the president, political conflicts between the president and parliament can easily spiral into political deadlock. The cost of repeated political deadlocks is the breakdown of democratic regime. The Weimar experience is another obvious case of a protracted power struggle between strong presidents and strong parliaments, causing democratic breakdown. Other than the grandstanding between the president and parliament, the support for different branches also plays a significant role. When citizen supportshave a coordination problem, it is hard to keep a check on powerful presidents with authoritarian tendency and it easily entices either branch to misjudge that they will win the battle if they opt for escalating the conflict further. Figure 1: Balanced Semi-presidentialism Strong President Strong Parliament Popular Support Popular Support

23 Second, nine countries: Belarus, Guine-Bissau, Haiti, Kyrgyzstan, Mali, Niger, Peru, Rwanda, and Russia (Putin) have either a strong president constitutionally or strong support for the president. It is worthy noting that Belarus, Kyrgyzstan, Peru and Russia have opted for electoral authoritarianism 22, especially when their powerful presidents gain high legitimacy from citizens, as shown in Figure 2. The dashed line with a single arrow connotes that the strong president prevails over the moderate parliament. The solid line with a single arrow indicates there is supportive relationship. In semi-presidentialism, normally the president wields considerable powers such as nominating prime minister and directing foreign and military affairs. When enforcing these powers, the president has to consider political dynamics in parliament or to consult the prime minister controlling a majority in parliament. In other words, the president is not unlimited to use his authority. However, there is one situation where the president s pen is mightier than any other president s pen in semi-presidential countries. 23 The president with unilateral decree power can promulgate policies to substitute legislation passed by the parliament. 24 Decree power is a proactive power for the president to be very influential on legislation. The powerful president justifies wielding this particular power in the name of accelerating reforms or solving crises. When an ambitious president faces an intransigent parliament, it is more possible to entice the president with strong authority to bypass it or use this unilateral power to execute policies. For example, the newest Russian constitution tilted power in favor of presidential power over parliamentary power and as a superpresidential regime it impeded the prospects of democratic stability during the presidency of Yeltsin (Clark, 1998; Fish, 2000). From , President Yeltsin had broadly used his decree power, which amounts to 1091 times, while 192 executive-initiated laws were only enacted (Remington et al, 1998, p. 310). Third, the remaining scenario is strong parliament in tandem with coordination problems. This can be characterized as a bipolarized parliament easily enticing regime breakdown. The effective number of political parties in Comoros is 2 and the number of 22 Regarding electoral authoritarianism, please see Schedler (2006). 23 The original usage of this metaphor is proposed to describe presidential decree power in Brazil. Related discussions can see Power (1998). 24 Decree power is defined as the authority of the executive to establish law in lieu of action by the assembly (Carey and Shugart, 1998, p.9). In addition, decree power can be classified into two categories. One is Constitutional Decree Authority (CDA) and the other is Delegated Decree Authority (DDA). Related discussions can see Carey and Shugart (1998). Here, it refers to constitutional decree authority.