The perils of semi-presidentialism: Confidence in political institutions in contemporary democracies

Abstract

The degree to which citizens perceive democratic political institutions as trustworthy indicates how well these institutions translate the citizenry’s interests into public policy and how effective and accountable they are seen to be. Low levels of public confidence in political institutions are an indicator of various political problems and are likely to raise concerns over democratic governance. Recent findings that trust in major political institutions has fallen over the last quarter of a century in many democracies have led scholars to examine individual and institutional factors associated with political confidence. Aiming to contribute to this burgeoning literature, this study investigates the impact of semi-presidentialism on public confidence in two major political institutions: the government and parliament. Testing our arguments in 29 democracies through a multilevel analysis, we have found that, compared to presidential and parliamentary systems, semi-presidentialism often generates dual-legitimacy problems, thereby reducing confidence in both government and parliament.

Keywords

Confidence semi-presidentialism accountability government parliament democracy

Introduction

The degree to which citizens perceive democratic political institutions as trustworthy indicates how well these institutions translate citizens’ interests into public policy and how effective and accountable they are seen to be (Hardin, 2000). Low levels of public confidence in political institutions have been found to be associated with various political problems, including government inefficiency (Steinmo, 1994), high levels of corruption in policymaking (Anderson and Tverdova, 2003), a negative evaluation of governments’ economic performance (Hetherington, 1998), socio-cultural problems such as crime and child poverty (Mansbridge, 1997) and illegal activities such as tax fraud (Marien and Hooghe, 2011). Declining confidence in political institutions can raise general concerns over the effectiveness of democratic governance (Berelson, 1952). If the decline is perceived as an outcome of the incumbent government’s poor performance, replacing that government in the next election can be seen as a potential remedy. However, if the outcome is a result of enduring constitutional rules, it is necessary to explain what factors lead to low levels of confidence in political institutions or boost the citizenry’s positive attitude toward them.

Despite their importance in determining public confidence in democracies, studies on public confidence have overlooked macro-level factors, with some notable exceptions (Anderson and Tverdova, 2001, 2003; Mishler and Rose, 1997; Newton, 2001; Newton and Norris, 2000). Many have noted that factors such as economic development, economic growth and a low level of corruption create a positive image of political institutions, thereby increasing confidence in them (Newton and Norris, 2000). Others have focused on how the communist legacy has created distrust in institutions (Howard, 2003; Mishler and Rose, 1997). However, most of these studies on political trust have overlooked developing countries, with few exceptions (e.g., Aydin and Cenker, 2012; Pharr et al., 2000).

While building upon pre-existing studies, this paper investigates the neglected impact of institutional design on public confidence in contemporary democracies. We argue that public confidence in political institutions is lower in semi-presidential systems, because they generate a higher likelihood of both dual-legitimacy concerns and intra-executive conflict. These issues lead to significant coordination, cooperation and accountability problems within the executive, which depresses citizens’ confidence in political institutions.

We test our arguments in 29 democracies using the fifth wave of the World Values Survey (WVS), conducted between 2004 and 2009. We find that public confidence is significantly lower in semi-presidential than in either presidential or parliamentary systems. However, we do not find a significant difference in the degree of confidence in government in parliamentary and presidential systems, although the former produces higher confidence scores in parliament than the latter.

Confidence in political institutions

Political trust—the degree to which people perceive that government is producing outcomes consistent with their expectations’ (Hetherington, 2005)¹—is an essential part of the link between citizens and political institutions. Thus, the efficiency and ability of policy-makers partly determines citizens’ confidence in them. Easton (1965) distinguishes between specific support for public authorities—the degree to which the public approves of government outputs or the performance of the political authorities—and diffuse support—the degree of satisfaction with the political system’s institutional structure. In this study, we concentrate on specific support for government and parliament.

How smoothly executive and legislative institutions function might play an important role in determining how much confidence individuals have in them. If an institutional arrangement generates inveterate political inefficiency, the government’s effectiveness in resolving problems may be reduced. If distrust in government is generated by institutional outcomes such as intra-executive conflict, a lack of accountability or inherent uncertainty, the government will eventually not be able to collect as many resources (e.g. taxation) to solve problems (Neustadt, 1990). This situation generates further public distrust in government institutions, and the replacement of a specific government in the next election will not necessarily increase confidence in the institutions if the gridlock is inherently institutional.²

Parliamentary, presidential and semi-presidential systems

Until Duverger’s (1980) seminal work on semi-presidentialism, the institutional relationship between the executive and legislative bodies was considered to be of two types: presidential and parliamentary. These two systems provide different forms of vertical accountability to voters and horizontal accountability between different political institutions. High accountability is expected to increase citizens’ involvement and thus their consent to the regime.

When a popularly elected president and a representative legislative chamber co-exist in a political system, the result is a divided executive structure, which can lead to questions regarding the legitimacy of and responsibility for policy decisions. Linz (1990) did not specifically discuss semi-presidentialism as a hybrid system in his seminal work on presidential systems, but some of the concerns he raised regarding presidential systems are also applicable to semi-presidentialism, including that of dual legitimacy (Linz, 1990). According to Linz, dual legitimacy produces confused voters and accountability problems (1994), and legitimacy problems arise in presidential systems when it is not clear who represents the real will of the people (1990).

Linz (1990) argued that parliamentary systems are much more flexible. In the event of irresolvable differences between the legislature and the executive, the former, using a vote of non-confidence, can remove the latter. In a presidential system, in contrast, conflict between the executive and the legislature is more persistent, resulting in poor performance. The winner-take-all nature of the presidential system can also lead to highly contested election results: ‘… the unsuccessful candidates may call into question the conduct of the election and the legitimacy of the President’s mandate, encouraging their supporters to take to the streets and overturn the result, and democracy, by force’ (Elgie, 2011:8).

General de Gaulle’s establishment of the Fifth Republic in France in 1958 to remedy the failure of political parties and coalition politics inspired Duverger’s (1980) seminal work, which discusses the distinct features of semi-presidentialism. Duverger (1980) defined semi-presidentialism as possessing three main characteristics: the president is directly elected, the presidential office has considerable constitutional authority and the cabinet and ministers are subject to legislative oversight. However, it is unclear whether or not a president has ‘considerable constitutional authority’ in a semi-presidential system (Elgie, 2009: 22). Instead, therefore, we have relied on Elgie’s (2009) definition, which avoids any reference to the observed behavior of political actors that could potentially raise an endogeneity problem. According to Elgie, a semi-presidential system is one in which there is a directly elected fixed-term president and the prime minister and cabinet are collectively responsible to the legislature (Elgie, 1999: 13).

We expect semi-presidentialism to exert a negative influence on citizens’ attitudes toward political institutions as a result of dual-legitimacy concerns and frequent intra-executive conflicts. To examine this hypothesis, it is first necessary to understand the two forms that the dual-legitimacy problem can take in semi-presidential systems, both of which have negative effects on the performance of political institutions and how citizens view them. The first is the problem of cohabitation (Elgie, 2011; Pierce, 1991). This refers to a political situation in which the president and the prime minister are members of or supported by rival parties and/or the legislature is opposed to the president and approves a prime minister and cabinet that are equally opposed to the president (Elgie, 2011: 12). The process of electing the prime minister and forming the cabinet is a conflict-ridden one requiring negotiations between various political actors, including the president and leading party members party (Schleiter and Morgan-Jones, 2009). In the process of forming a government, conflicts between political agents may lead to various actions, including dissolving the assembly and initiating judicial review (Strøm et al., 2003: 673–681). Shugart and Carey (1992) argue that the possibility of a divided executive is built into semi-presidentialism. They stress the perils of cohabitation and argue that it leads to regime crisis, especially in younger democracies. They suggest that, in a semi-presidential system, the constitution should be clear on the specific responsibilities of and differences between the two executives (Shugart and Carey, 1992). Elgie (2010) acknowledges that there is lack of empirical evidence to support the claim that cohabitation leads to regime collapse in electoral democracies, but he argues that cohabitation can provoke regime crises in younger democracies, and that semi-presidentialism is a perilous constitutional design in such settings.³

Linz and Stepan (2011) examined the second form of the dual-legitimacy problem in semi-presidentialism, conflict between the executive and the legislature, and argued that it may weaken the legitimacy of a political system. A political vacuum can occur in a highly fragmented legislature, which may be filled by executive decree, thereby intensifying the conflict and reducing confidence in institutions. In semi-presidentialism, power-sharing between executive bodies precludes a clear separation of powers, which often leads to constitutional ambiguities (Skach, 2007). Ostrom (2005) argues that the division of powers between the executive bodies must be completely and consistently delineated. De Raadt (2009) argues that in the Polish semi-presidential system, ambiguity leads to increased conflict because of the substantial room to maneuver for political actors, which makes it especially difficult for citizens to determine who to hold accountable.

In short, the sharing of executive powers by the president and the prime minister makes it more difficult to establish executive coordination, which affects the performance of political institutions. To be clear, we are not claiming that dual-legitimacy problems are specific to semi-presidentialism or that hybrid systems necessarily result in worse political performance. It may be that in certain situations dual legitimacy promotes outcomes that citizens prefer or contributes to the ability to reach compromises or coordinate policy, as is the case in France (Elgie, 2001, 2011; Leuffen, 2009). But this power-sharing arrangement could be detrimental for transitional democracies, as Kirschke (2007) found in her analysis of neopatrimonial states in Africa, where the viability of government hinges upon how well the executive is able to reward its key supporters. In a recent work, Kim (2014) has found that the absence of checks on the presidential authority is the primary cause of poor democratic outcomes in semi-presidentialism. We abstain from building deterministic arguments regarding the relationship between semi-presidentialism and accountability; however, considerable research has found that semi-presidentialism produces greater intra-executive conflict and accountability concerns than other systems (Linz, 1994; Skach, 2007). Even though semi-presidentialism in France is regarded as a model of success for transitioning democracies in Europe, we approach this view with caution. In the French case, the role of the prime minister is enhanced in times of cohabitation, and at other times prime ministers are subordinated to the president. The division of responsibility within the executive varies with the personalities of those who hold office and the conditions under which they serve (Schain, 2008). The French system operates with relative stability because it has been able to provide either a presidential majority or a majority of the opposition party in the General Assembly, which enables that party to govern (Stepan and Suleiman, 2001). However, ambiguities over executive power are still intrinsic to the semi-presidential system and generate distrust of government and politics in general (Schain, 2008).

Subtypes of the semi-presidential system may lead to different policy outcomes and a differing likelihood of executive conflict (Raunio, 2012; Roper, 2002; Tsai, 2008).⁴ Even so, the dual-legitimacy problem tends to occur in all subtypes of this system, whether the president can only govern with the support of the legislature or the government is responsible to the president and parliament (Elgie, 2011). With fewer incentives for cooperation, semi-presidentialism—especially when the government is responsible to the president and parliament—tends to influence democratic performance negatively. Linz (1994: 55) noted that semi-presidentialism is subject to ‘instability and inefficiency’ even in the most favorable political circumstances—when the president, prime minister, and ministers belong to the same party, which has a legislative majority—because the policies produced by the president will necessarily differ from those of the prime minister. Furthermore, ministers will appeal to the president when their proposal is not well received by the prime minister, which may delay decision-making and result in inefficient policies (Elgie, 2011).

Several examples from semi-presidentialism support this claim. For instance, semi-presidentialism has existed in Taiwan since the constitutional change in 1997. Since then, the political system has experienced both unified and divided governments—characterized by the presence and absence of majority support for the president in the Legislative Yuan, respectively. The Taiwanese constitution clearly states that the president appoints the premier without the consent of the Legislative Yuan, which increases the powers of the president vis-à-vis the premier and the legislature. Even constitutional scholars disagree as to whether the president can unilaterally remove the premier, as Matsumoto (2013) has demonstrated. This ambiguity over the responsibilities of the president makes it difficult for citizens to determine whom to hold accountable for policy-making processes. The decision-making process becomes dependent on individual factors, such as the president’s desire to exert power and his popularity within his own party, as well as contingent factors, such as the percentage of seats held by the president’s party in the legislature and party discipline within the legislative parties (Matsumoto, 2013). Additionally, in votes of non-confidence, the president also has the right to dissolve parliament instead of replacing the premier, which adds further uncertainty over accountability.

In addition to this constitutional ambiguity, the actual functioning of the political system in Taiwan also depends in part on the political circumstances. Research into the approval rates of legislation generated by the president in Legislative Yuan and the average number of days spent deliberating bills before they pass reveals that presidential power is more efficient when there is a unified government (Sheng, 2003; Matsumoto, 2013). We argue that these kinds of fluctuations are inherent to semi-presidential systems in which the constitutional powers of executive and legislative political actors are imperfectly defined and the relationship between them is contingent upon political conditions. We do not deny that semi-presidentialism can create situations in which political accountability is higher. In practice, the political system in Taiwan operated similarly to a presidential one until 2000, while Lee Teng-hui from the Kuomintang Party held the presidency and a legislative majority. In 2000, however, the Democratic Progressive Party candidate, Chen Sui-bian, won the presidential elections. Between 2000 and 2006, Taiwan experienced a divided government in which the president belonged to a party that possessed less than half the seats in parliament, resulting in gridlock between 2000 and 2007 (Huang, 2006). According to Wu’s (2009) findings, the Taiwanese public blames the divided government structure in Taiwan for the poor conduct of political affairs throughout this period. This period of gridlock in Taiwanese politics, we argue, explains the decline in confidence in governments between 2001 and 2007, which is also reflected in East-Asia Barometer survey results (Asia Barometer, 2014). The survey reveals that 47% of the population distrusted the government in 2001, and 53% in 2007. During the same period, distrust in parliament also increased from 68% to 71%. As a result of Taiwan’s semi-presidential system, not only is public confidence in national political institutions continuously at low levels, but the public is also able to criticize and thus castigate national political institutions for political crises because of the mutual blaming between the president and the premier during such crises.

In Portugal, constitutional changes in 1982 significantly reduced presidential power. The president at the time, António Ramalho Eanes, raised his concern over the accountability of the political system. He claimed that with the new constitution in 1982, it became difficult to tell who held certain responsibilities (Bruneau and Macleod, 1986: 128). Major changes to the constitution included the restriction of the president’s ability to dissolve the cabinet and the Assembly of the Republic, the parliament, and restrictions on the president’s ability to veto parliamentary legislation. As a result of these changes, the cabinet became politically responsible to the Assembly rather than the president (Neto and Lobo, 2012).

However, despite the restrictions on presidential power, the 1982 Constitution granted the president some tools to influence the government’s actions, including the power to manipulate ministerial appointments and a complicated veto power. Owing to these formal rules, Eanes was strong enough to use a threat to dissolve parliament to compel Prime Minister Mário Soares to refrain from repeatedly shuffling the cabinet. He also urged the prime minister to make concessions to the members of his coalition government (Bruneau and Macleod, 1986: 141). The Portuguese president’s powers regarding ministerial appointments differ under majority and minority governments. Under majority governments, the president has less influence. At times of cohabitation, presidents prefer non-partisan ministers in the cabinet. Between 1976 and 2004, 28% of cabinets were non-partisan in minority-government situations and 17% in majority governments (Neto and Strøm, 2006). During this period, presidents used their veto power 62 times (Neto and Lobo, 2012). Thus, the influence of the president is conditional not only on the formal rules in the constitution, but also on the nature of government and the president’s partisanship.

In addition to these formal powers, we see examples of Portuguese presidents using public communication to put pressure on governments, which in 1992 resulted in a closed-door fight between and public scrutiny of the two parts of the executive when Prime Minister Aníbal Cavacio Silva stated that he would not permit an influencing president to turn into an interfering one (Neto and Lobo, 2012: 63).

These examples from Taiwan and Portugal reinforce our arguments regarding the vulnerability and poor performance of the semi-presidential system and its role in plunging political systems into frequent gridlock. In addition, they demonstrate the difficulty of holding executives in a semi-presidential system accountable for poor conduct during such crises. These concerns over accountability decrease public trust in governments during such crises, which contribute to the fragility of the semi-presidential system, in which low public confidence becomes endemic.

In accordance with the theoretical and empirical discussion, we offer the following hypothesis:

Hypothesis: Confidence in political institutions will be lower in semi-presidential systems than in presidential and parliamentary systems.

Before conducting a rigorous analysis of the determinants of political confidence, we display the relationship between political systems and political trust in democratic countries. Figures 1(a) and 1(b) show that the mean scores of confidence in government and parliament are lower in semi-presidential systems than in the other two systems. Parliamentary systems produce better results for confidence in parliament, while presidential ones do well for confidence in government.⁵ But semi-presidentialism results in a worse overall outcome for confidence in both parliament and government. These preliminary findings supporting our expectations regarding semi-presidentialism require more scrutiny, which we turn to in the multivariate-analysis section.

Figure 1a.

Confidence in government.

Figure 1b.

Confidence in parliament.

Research design

Data and methods

This study uses Wave 5 of the WVS, which was conducted in 57 countries. We relied on the Polity IV of the survey year to determine whether the political regime is democratic or not.⁶ Polity IV includes constructed annual measures for both institutionalized democracy and autocracy (Marshall et al., 2013). In line with other studies that base their categorization of regimes on the Polity IV score, we considered countries to be democratic, and added them to the analysis, if they scored 6 or higher (e.g. Howard, 2003; Reed, 2000). Due to our limiting of the analysis to democracies and missing observations, our sample size decreases to 23,115 individual observations and 29 countries.

Dependent variables

The WVS asks how much confidence one has in government and parliament. The respondents have four options, specifically a great deal and quite a lot on the positive end and not very much and not at all on the negative. We run a t-test to determine whether the level of support for parliament and government among respondents is identical or not. The results confirm that confidence in parliament and government are generated through different processes.⁷ In sum, there are two dependent variables that measure (a) confidence in government and (b) confidence in parliament. The scores for each individual range between 0 and 3, depending on the four possible responses in the survey, where higher numbers are associated with higher confidence in political institutions.

Main independent variables

Presidential, parliamentary and semi-presidential systems

One of our main independent variables represents the relationship between the executive and legislative branches. This variable is a categorical one with three values. The countries in our dataset are categorized as Presidential, Parliamentary, or Semi-presidential systems. We made our decisions on the categorization of regimes by using Bormann and Golder (2013) and performed robustness checks by looking into other sources (Cheibub, 2007; Duverger, 1980; Elgie, 2011).

National-level political and economic factors

We considered three important political factors that may affect the level of confidence in institutions in a given country.⁸ We expect that citizens of new democracies, those that have made a transition to democracy in the post-war era, will have less confidence in institutions.⁹ We also use the longevity of democracy in a country to control its effect (Howard, 2003). Kim (2014) found that checks on presidential power are likely to increase regime performance. In order to control for this possibility, we utilized a variable that measures the relative strength of parliament over the executive derived from Fish and Kroenig (2009).

The macro-economic conditions of a country affect how citizens perceive political institutions. As such, positive developments, such as high economic growth and low unemployment, are expected to increase the positive perception of governments.¹⁰ Moreover, this perception may carry over to citizens’ perception of other domestic political institutions. Economic development and unemployment data come from the World Bank’s World Development Indicators.¹¹ In order to test the effect of inequality on citizens’ perceptions, we use the standardized Gini index (Solt, 2009). The Gini index has a theoretical range from 0 to 100; the higher the number, the greater the inequality.¹²

Individual-level factors

Winner: While all individuals have an opinion on a government’s performance, their possible allegiance to the government can play a role in how they perceive its actions (Anderson and Tverdova, 2001, 2003).¹³ One expects that those who voted for the parties in power are more likely to have higher levels of confidence. Those who voted in the last election for any party within the government, including coalition partners, have been coded as 1, and 0 otherwise.

Voter: The survey asks whether or not the respondent will vote in the next election. We expect that nonvoters will have less confidence in both executive and legislative institutions (Anderson and Tverdova, 2003). Those who professed that they would vote for the incumbents in the upcoming elections are coded as 1, otherwise 0.

Social Capital: A decline in confidence in government partly stems from a decline in social capital, which suggests that social capital increases political trust (Putnam, 1993). Hardin also argues that participation in organizations especially affects views regarding the trustworthiness of political institutions (Hardin, 2000). Following Putnam and Hardin, we used respondents’ membership in civil-society organizations.¹⁴ The WVS asks respondents whether they are an active or passive member of eight types of civil-society organizations (environmental organization; humanitarian or charitable organization; consumer organization; professional association; religious organization; art, music or educational organization; sport or recreational organization; labor union and others). Those who are active members of civil-society organizations are coded as 1, while passive members and nonmembers are coded as 0. We sum up the answers for all associations and create an indexed variable that ranges from 0 to 8.

Political Interest: It is expected that individuals who are more interested in politics hold higher confidence in institutions. This variable asks how important politics is to the respondent, and the answers can range from 0 (not very important at all) to 3 (very important).

Left-Right Scale: According to the literature on partisanship, right-wing voters are more likely to have a high degree of confidence in political institutions (Listhaug and Wiberg, 1995). Thus, the left/right scale was introduced to the models as a control variable. Higher numbers are associated with right-wing individuals.

We also expect that satisfaction with one’s household income and religiosity increase citizens’ confidence in political institutions (Catterberg and Moreno, 2006). The former is measured by the degree to which a respondent views his or her current household income as satisfactory. The variable ranges from 1 to 10. Self-reported religiosity is a variable measured on a scale of 0 (atheist) to 2 (very religious).

Other Control Variables: Finally, the other control variables included are the respondent’s education, age and sex (coded 1 for men), built on previous studies on the relationship (Verba et al., 1995). The impact of education on political trust is controversial. Almond and Verba (1963) see it as positively correlated; Dalton (1996), in contrast, asserts that high levels of education lead to critical attitudes toward political institutions; and Listhaug and Wiberg (1995) argue that its impact varies depending on the institution in question.

Methodology

The Ordinary Least Squares method (OLS) assumes that individual-level errors are uncorrelated with others in a given country, which causes a Type I error (Steenbergen and Jones, 2002). In order to account for the relationship between individual- and country-level factors, we rely on the multilevel model. The statistical analyses include individual- and contextual-level data and models.¹⁵ The first model that includes individual-level variables is:

Confidence in Parliament / Governmen t_{=} β_{0 j} {+ β}_{1 j} {Winner}_{ij} {+ β}_{2 j} {Voter}_{ij} {+ β}_{3 j} {Social Capital}_{ij} {+ β}_{4 j} {Religiosity}_{ij} {\dots β}_{kj} x_{kij} {+ e}_{ij .}

The introduction of country-level variables to measure their independent effects on confidence in parliament and government is as follows:

β_{0 j} {= γ}_{00} {+ γ}_{01} {Parliamentarism}_{j} {+ γ}_{02} {Semi - presidentialism}_{j} {+ γ}_{03} {Development}_{j} {+ γ}_{04} {Unemployment}_{j} {+ γ}_{05} {Inequality}_{j} & & .. {+ U}_{0 j}

Results

We first direct our attention to individual-level survey responses before detecting the impact of macro-level variables. Model 1 in Table 1 indicates that citizens who professed their intention to vote in the next elections have higher confidence in government. Being a non-voter implies that they are dissatisfied with the existing political parties and do not expect good policy outcomes from governments. This expectation is reinforced by the fact that respondents who self-declare that they are interested in politics also have more confidence in government.

Table 1.

Individual determinants of confidence in political institutions.

	Confidence in Government		Confidence in Parliament
	Model 1	Model 2	Model 3	Model 4
Winner	–	0.2411 ***	–	0.1752 ***
Winner		0.0759		0.0387
Voter	0.2864 ***	0.1775 ***	0.2707 ***	0.1916 ***
Voter	0.0385	0.0432	0.0317	0.0323
Social Capital	0.0188 ***	0.0191 ***	0.0065	0.0067
Social Capital	0.0052	0.0048	0.0072	0.0071
Religiosity	0.0760 ***	0.0649 ***	0.0681 ***	0.0063 ***
Religiosity	0.0192	0.0160	0.0195	0.0169
Household Finance	0.0338 ***	0.0323 ***	0.0311 ***	0.0300 ***
Household Finance	0.0042	0.0041	0.0046	0.0044
Ideology	0.0051	0.0062	0.0056	0.0063
Ideology	0.0137	0.0127	0.0077	0.0069
Education	−0.0172 **	−0.0153 **	−0.0066	−0.0052
Education	0.0070	0.0070	0.0062	0.0061
Female	0.0175	0.0173	−0.0070	−0.0070
Female	0.0151	0.0146	0.0168	0.0164
Age	0.0009 *	0.0007	0.0007	0.0005
Age	0.0005	0.0006	0.0005	0.0005
Political Interest	0.0928 ***	0.0926 ***	0.1044 ***	0.1042 ***
Political Interest	0.0132	0.0125	0.0142	0.0139
Intercept	1.3699 ***	1.3622 ***	1.2082 ***	1.2048 ***
Intercept	0.0428	0.0423	0.0553	0.0542
Random Effect
Intercept	0.0537	0.0533	0.0913	0.0877
Level-1	0.6346	0.6232	0.5627	0.5567

Note: Full Maximum Likelihood is used with HLM 6.0; entries are unstandardized coefficients with robust standard errors. N. 23115, *p < .1; **p< 0.05 and ***p< 0.01 (two–tailed tests).

As expected, individuals who claim to be more religious, and those who score higher on social capital are more likely to have confidence in government, whereas highly educated individuals score the opposite as they form critical views of government. Meanwhile, those who believe that politics is important show more confidence in government. The control variables, such as ideology, age and gender, do not have statistically significant coefficients.

In Model 2, we added a winner variable to our model. This model shows that supporters of the governing party are more likely to have confidence in the government. This supports the idea that when governments are representative of larger groups’ interests, they are more likely to generate trust among citizens, which is crucial for a well-functioning political system.

Models 3 and 4 run the same models for confidence in parliament. Model 3 differs from Model 1 in two ways. First, the social-capital variable is not statistically significant enough to account for individuals’ confidence in parliament. Another difference is that educated individuals are less critical of parliament than they are of government. When we add the winner variable to Model 4, the results are similar to its relationship with confidence in government. The main finding is that being supportive of the government has a positive effect on individuals’ confidence in political institutions.

Table 2 depicts the determinants of confidence in government in a multilevel analysis that incorporates individual-level variables and country-level economic and institutional variables. According to Model 1 in Table 2, compared to presidential systems (the reference category), the semi-presidential system produces less confidence in government. Note that when we make the parliamentary system our reference category for all models, we end up observing the negative impact of the semi-presidential system on confidence in political institutions. We find that parliamentary systems are not significantly different from presidential systems in producing confidence in governments. Taking into account issues that have been widely discussed in the existing literature, in Model 2 we control for national-level political and economic variables such as the relative strength of the parliament vis-à-vis the executive, the age of the democracy, new versus long-standing democracies, economic development, inequality and unemployment. Our national-level control variables yield no statistical significance, suggesting that none of the first three factors have a strong effect on one’s confidence in government. When we control for macro-economic variables in Model 2, the relationship between semi-presidentialism and confidence in government still holds. Economic development has a statistically significant coefficient—but, surprisingly, a negative one. However, a substantive interpretation of the coefficient suggests that its impact is very small; every $1,000 increase in gross domestic product (GDP) per capita is associated with an increase of 0.007 in confidence in government.

Table 2.

Semi-presidentialism, parliamentarism and confidence in government and parliament.

	Confidence in Government		Confidence in Parliament
	Model 1	Model 2	Model 3	Model 4
Main Variables
Parliamentarism	−0.1122	−0.0572	0.2373 *	0.2391 *
Parliamentarism	0.1002	0.1240	0.1212	0.1363
Semi-presidentialism	−0.3048 ***	−0.3634 ***	−0.1696 *	−0.3019 **
Semi-presidentialism	0.0854	0.1214	0.0701	0.1348
National-level Control Variables
Strength of Parliament	–	−0.4020	–	−0.7021
Strength of Parliament		0.6652		0.7477
New Democracy	–	−0.0529	–	−0.0393
New Democracy		0.1583		0.1705
Age of Democracy	–	0.0004	–	0.0030
Age of Democracy		0.0024		0.0029
Development	–	−0.0071 *	–	−0.0134
Development		0.0031		0.0086
Inequality	–	−0.0073	–	−0.0152
Inequality		0.0071		0.0091
Unemployment	–	−0.0040	–	0.0085
Unemployment		0.0080		0.0107
Economic Growth	–	−0.0037	–	−0.0085
Economic Growth		0.0299		0.0373
Intercept	1.3667 ***	1.3623 ***	1.2047 ***	1.2049 ***
Intercept	0.0428	0.0327	0.0446	0.0398
Random Effect
Intercept	0.0537	0.0466	0.0634	0.0689
Level-1	0.6346	0.6232	0.5567	0.5567

Note: Full Maximum Likelihood is used with HLM 6.0; entries are unstandardized coefficients with robust standard errors. *p <0 .1; **p<0 0.05 robust standard errors. *p <0.1; **p< 0.05 category. N:23115.

One may claim that reduced confidence in government is generated by contingent factors, such as cohabitation or coalitional governments, instead of constitutional set-up. Yet, when we analyzed the level of confidence in government for the seven cases of semi-presidentialism in our sample by tracing their average scores of confidence in earlier rounds of the WVS, we found no clear evidence that periods of cohabitation generate a sufficiently large difference in the outcome compared to periods of unitary government.¹⁶ When we conducted the multilevel analyses by including a binary control variable of coalition versus single-party governments, our results remained robust. This finding suggests that the difference between semi-presidentialism and the two other systems goes beyond differences in the contingent factors within semi-presidentialism. Additionally, to control for the popularity of the government or occasional political crises that might distort the impact of institutional factors on political trust, we added a variable for governments’ honeymoon period, measured by the months in office, and a variable for political crises, measured by Arthur Banks’ dataset. None of these factors change the negative impact of semi-presidentialism on confidence in government.¹⁷

When we analyze confidence in parliament, we observe similar results in Models 3 and 4 in Table 2. No national-level political or economic factor has any statistically significant effect on confidence in institutions. When we analyze the political-system variables, we observe that semi-presidentialism produces the lowest confidence in parliament. Therefore, confidence in parliament also suffers in semi-presidential systems. Another important finding is that, compared to the presidential system, the parliamentary system tends to produce more confidence in parliament.

In accordance with our argument, the effect of political systems should only be applied to the institutions held accountable by the public, and not to those that are not elected by the public. To test whether this assumption is correct, we ran our model for confidence in another component of the political system, namely the judiciary, because it is not directly accountable to the public. We found that semi-presidentialism has no statistically significant effect on confidence in the judiciary, which further confirms our explanation by focusing on the impact of accountability.¹⁸

Robustness check

In accordance with our argument, we expected that confidence in governments and parliaments would decline over time in countries with semi-presidential systems. Using the means of difference for Round 3 (1995–1998) to Round 5 (2005–2009) of the WVS for the 22 cases for which data are available in both rounds, we tested the extent to which countries with different systems experienced changes in confidence in institutions over time.¹⁹

First, there is no evidence that semi-presidential systems are chosen in countries with low political trust. Even though semi-presidential systems scored higher than others in confidence in government in Round 3 by averaging 1.39, these countries had the lowest levels of confidence in government in Round 5, with a score of 1.06. The mean score for confidence in parliament also declined from 1.22 to 0.83 within the same period. In parliamentary systems, there was an increase in support for parliament and government between two rounds. The results are mixed for presidential systems: consistently with our earlier results, they provide a worse outcome only for parliaments. Consequently, these results provide us with empirical evidence that the lower confidence in semi-presidential systems is a result of a decline in countries that have maintained semi-presidential systems between Rounds 3 and 5 of the WVS.

Given the fact that, in our sample of 29 contemporary democracies, all cases of semi-presidentialism except one are from post-communist countries and the conventional wisdom that the communist legacy produces lower political trust in transitional periods, one may raise concerns over the potentially spurious relationship between regime-legacy and political-institution variables to explain political trust in governments and parliaments.

To eliminate this possibility, we executed a difference of means test only on post-communist countries by enlarging the data to include all available WVS rounds (Rounds 2 to 6).²⁰ This also enables us to reduce unobserved heterogeneity due to the region-specific factors in our models. Running the difference of means test between semi-presidential and non-semi-presidential systems in 19 post-communist countries with 35 observations, we found that average public confidence in government and parliament is significantly lower in semi-presidentialism.

Conclusion

Linz (1990) established the dichotomy between parliamentary and presidential systems and pointed out that presidential systems are prone to conflict because of their dual-executive structure. In this dichotomy, semi-presidentialism, by definition, appeared as a hybrid regime between two major systems and was expected to produce mediocre performance. When Charles de Gaulle initiated the French Fifth Republic in 1958 as an attempt to remedy the governmental instability of the fourth Republic (that had been created after World War II), semi-presidentialism was considered as a balanced constitutional design. The semi-presidential system gained credit for France’s relative political stability since then, and has inspired many countries experiencing democratic transition.²¹ However, Duverger’s (1980) and Elgie’s (2001) later conceptualization of semi-presidentialism demonstrated the distinctive features of this system and promoted scholarly attention towards detecting its political outcomes. In this respect, our study adds to studies that examine the impact of semi-presidentialism on political behavior (Chang, 2014; Elgie, 2011; Skach, 2007). In addition, several case studies on transitional democracies and theoretical discussions have revealed that semi-presidentialism is a risky constitutional design. Our analysis of the newest transitional democracies, namely post-communist countries, shows that the semi-presidential system is conducive to lower public confidence in political institutions than presidential and parliamentary systems.

Finally, and most importantly, this study highlights the importance of bridging two seemingly distinct literatures, namely those on political systems and confidence in political institutions. Although earlier studies have investigated the impact of corruption, the post-communist legacy, economic performance and cultural factors on confidence in political institutions, they have neglected the effect of the executive structure. Our findings also have implications for democratization studies, such as those by Linz (1990) and Stepan and Skach (1993), as the choice of institutions is crucial to the consolidation of new democracies.

Our findings reveal that parliamentary systems produce higher levels of confidence in legislative institutions than do presidential ones, but there is no significant difference between them on confidence in government. We caution the reader on two shortcomings of this study. First, it is limited by the lack of longitudinal data on all contemporary democracies. As a consequence, future studies should test whether the difference in political regimes’ impact on political confidence is conclusive or not. Second, and also due to the limits of the data, we cannot test the degree to which subtypes of the semi-presidential system (premier-presidentialism and president-parliamentarism) have different impacts on the degree of political confidence. Nevertheless, the available data on the 29 countries studied here suggest that the impact of semi-presidentialism on political trust is distinct from that of presidentialism and parliamentarism—and that that impact is a negative one.

Footnotes

Acknowledgements

We would like to thank Cerem I Cenker-Ozek, H Ege Ozen, Sema Akboğa, Daniel Moure, the anonymous reviewers and Mark Kesselman for their comments and suggestions, and the participants at the Empirical Studies in Political Analysis Workshop in Antalya, Turkey, in January 2015, for their feedback.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Notes

Author biographies

Yüksel Alper Ecevit received his Ph.D. in Political Science from Binghamton University in 2013. He joined Bahcesehir University in the fall of 2013. His research interests are the politics of enfranchisement, political institutions and immigration politics.

Ekrem Karakoç is an assistant professor at the Department of Political Science at Binghamton University, SUNY. He specializes in Comparative Politics with a focus on political institutions, political behavior and democratization.

References

Almond

Gabriel

Verba

Sidney

(1963) The Civic Culture: Political Attitudes and Democracy in Five Nations. Princeton: Princeton University Press.

Anderson

Christopher Johannes

Tverdova

Yuliya

(2001) Winners, Losers, and Attitudes About Government in Contemporary Democracies. International Political Science Review 22(4): 321–338.

Anderson

Christopher Johannes

Tverdova

Yuliya

(2003) Corruption, Political Allegiances, and Attitudes Toward Government in Contemporary Democracies. American Journal of Political Science 47(1): 91–109.

Asia Barometer (2014) Survey Results are available at : http://www.asianbarometer.org/newenglish/surveys/ (accessed 21 December 2014).

Aydın

Aylin

Cenker

Cerem Işıl

(2012) Public Confidence in Government: Empirical implications from a developing democracy. International Political Science Review 33(2): 230–250.

Berelson

Bernard

(1952) Democratic Theory and Public Opinion. Public Opinion Quarterly 16: 313–30.

Bormann

Nils-Christian

Golder

Matt

(2013) Democratic Electoral Systems Around the World, 1946–2011. Electoral Studies, 32(2): 360–369.

Bruneau

Thomas Charles

Macleod

Alexandre

(1986) Politics in Contemporary Portugal: Parties and the Consolidation of Democracy. Boulder: L Rienner Publishers.

Catterberg

Gabriela

Moreno

Alejandro

(2006) The Individual Bases of Political Trust: Trends in new and established democracies. International Journal of Public Opinion Research 18(1): 31–48.

10.

Chang

Chun-Hao

(2014) Cohabitation in Semi-Presidential Countries. Social Sciences 3.2 (2014): 31–43.

11.

Cheibub

Jose Antonio

(2007) Presidentialism, Parliamentarism, and Democracy. Cambridge: Cambridge University Press.

12.

Cook

Timothy

Gronke

Paul

(2005) The skeptical American: Revisiting the meanings of trust in government and confidence in institutions. Journal of Politics 67(3): 784–803.

13.

Dalton

Russell

(1996) Citizen politics: Public Opinion and Political Parties in Advanced Industrial Democracies. Chatham, NJ: Chatham House.

14.

De Raadt

Jasper

(2009) Contestable constitutions: Ambiguity, conflict, and change in East Central European dual executive systems. Communist and Post-Communist Studies 42(1): 83–101.

15.

Duverger

Maurice

(1980) A new Political System Model: Semi-presidential government. European Journal of Political Research 8(2): 165–187.

16.

Easton

David

(1965) A System Analysis of Political Life. New York: Wiley-Blackwell

17.

Easton

David

(1975) A Re-Assessment of the Concept of Political Support. British Journal of Political Science 5(4): 435–457.

18.

Elgie

Robert

(ed.) (1999) Semi-presidentialism in Europe. Oxford: Oxford University Press.

19.

Elgie

Robert

(2001) Divided Government in Comparative Perspective. Oxford: Oxford University Press

20.

Elgie

Robert

(2009) Duverger, Semi-Presidentialism and the Supposed French Archetype. West European Politics 32(2): 248–267.

21.

Elgie

Robert

(2010) Semi-Presidentialism, Cohabitation and the Collapse of Electoral Democracies, 1990–2008. Government and Opposition 45(1): 29–49.

22.

Elgie

Robert

(2011) Semi-Presidentialism: Sub-Types and Democratic Performance. Oxford: Oxford University Press.

23.

Fish

Stephen

Kroenig

Matthew

(2009) The Handbook of National Legislatures: A Global Survey. New York: Cambridge University Press.

24.

Hardin

Russell

(2000) The Public Trust. In Pharr

Susan J

Putnam

Robert D

(eds), Disaffected Democracies: What’s Troubling the Trilateral Countries? Princeton: Princeton University Press, 31–51.

25.

Hetherington

Marc

(1998) The Political Relevance of Political Trust. American Political Science Review 92: 791–80.

26.

Hetherington

Marc

(2005) Why Trust Matters: Declining Political Trust and the Demise of American Liberalism. Princeton, NJ: Princeton University Press.

27.

Howard

Marc Morje

(2003) The Weakness of Civil Society in Post-communist Europe. Cambridge University Press.

28.

Huang

Thomas Weishing

(2006) The President Refuses to Cohabit: Semi presidentialism in Taiwan. Pacific Rim Law and Policy Journal 15: 375–402.

29.

Kim

Young Hun

(2014) A troubled marriage? Divided Minority government, cohabitation, presidential powers, president-parliamentarism and semi-presidentialsm. Government and Opposition Available on CJO 2014 doi:10.1017/gov.2014.23

30.

Kirschke

Linda

(2007) Semipresidentialism and the Perils of Power-Sharing in Neopatrimonial States. Comparative Political Studies 40(11): 1372–1394.

31.

Lawrence

Robert

(1997) Is it really the economy, stupid? In Zelikow

Philip

Nye

Joseph

King

David C

(eds) Why People Don’t Trust Government. Cambridge: Harvard University Press, 111–132.

32.

Leuffen

Dirk

(2009) Does Cohabitation Matter? French European Policy-Making in the Context of Divided Government. West European Politics 32(6): 1140–1160.

33.

Linz

Juan Jose

(1990) The Perils of Presidentialism. Journal of Democracy 1(1): 51–69.

34.

Linz

Juan Jose

(1994) Presidential or Parliamentary Democracy: Does it make a difference? In Linz

Jose

Valenzuela

Arturo

(eds) The Failure of Presidential Democracy. Baltimore: The Johns Hopkins Press.

35.

Linz

Juan Jose

Stepan

Alfred

(2011) Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post-communist Europe. JHU Press.

36.

Listhaug

Ola

Wiberg

Mati

(1995) Confidence in political and private institutions. In Klingemann

Hans-Dieter

Fuchs

Dieter

(eds) Citizens and the State. Oxford: Oxford University Press: 298–322.

37.

Mansbridge

Jane

(1997) Social and cultural causes of dissatisfaction with U.S. Government. In King

David C

Nye

Joseph S

Zelikow

Philip D

(eds) Why People Don’t Trust Government Cambridge, MA: Harvard University Press.

38.

Marien

Sophie

Hooghe

Marc

(2011) Does Political Trust Matter? An Empirical Investigation into the Relation Between Political Trust and Support for Law Compliance. European Journal of Political Research 50(2): 267–291.

39.

Marshall

Monty

Gurr

Ted Robert

Jaggers

Keith

(2013) Polity IV Project: Political Regime Characteristics and Transitions, 1800–2012. Centre for Systemic Peace.

40.

Matsumoto

Mitsutoyo

(2013) Presidential strength and party leadership in Taiwan. In Kasuya

Yuko

(ed) Presidents, Assemblies and Policy-making in Asia. Palgrave Macmillan, 107–134.

41.

Mishler

William

Rose

Richard

(1997) Trust, Distrust and Skepticism: Popular evaluations of civil and political institutions in post-communist societies. Journal of Politics 59: 418–451.

42.

Neto

Octavio Amorim

Lobo

Marina Costa

(2012) Portugal’s semi-presidentialism (Re) considered: An Assessment of the President’s Role in the Policy Process, 1976–2006. In Royo

Sebastian

(ed) Portugal in the Twenty-First Century: Politics, Society and Economics, Lanham, MD: Lexington Books, 49–69.

43.

Neto

Octavio Amorim

Strom

Kaare

(2006) Breaking the Parliamentary Chain of Delegation: Presidents and non-partisan cabinet members in European democracies. British Journal of Political Science 36(4): 619.

44.

Newton

Ken

Norris

Pippa

(2000) Confidence in public institutions. In Pharr

Susan J.

Putnam

Robert

(eds) Disaffected Democracies. Princeton, NJ: Princeton University Press.

45.

Newton

Ken

(2001) Trust, Social Capital, Civil Society, and Democracy. International Political Science Review 22(2): 201–214.

46.

Neustadt

Richard Elliot

(1990) Presidential Power. New York: Macmillan.

47.

Ostrom

Elinor

(2005) Understanding Institutional Diversity. Princeton University Press.

48.

Pharr

Susan

Putnam

Robert

Dalton

Russell

(2000) A Quarter-Century of Declining Confidence. Journal of Democracy 11(2): 5–25

49.

Pierce

Roy

(1991) The Executive Divided Against Itself: Cohabitation in France, 1986–1988. Governance 4(3): 270–294.

50.

Putnam

Robert

(1993) Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press.

51.

Raunio

Tapio

(2012) Semi-Presidentialism and European Integration: Lessons from Finland for constitutional design. Journal of European Public Policy 19(4): 567–584.

52.

Reed

William

(2000) A Unified Statistical Model of Conflict Onset and Escalation. American Journal of Political Science 44(1): 84–93.

53.

Roper

Steven

(2002) Are all Semipresidential Regimes the Same? A Comparison of Premier-Presidential Regimes. Comparative Politics 34(3): 253–272.

54.

Schain

Martin

(2008) Politics in France. In Almond

Gabriel

Powell

G Bingham

Dalton

Russell J

Strom

Kaare

(eds) Comparative Politics Today, 9th edition. Pearson Education.

55.

Schleiter

Petra

Morgan-Jones

Edward

(2009) Party Government in Europe? Parliamentary and Semi-Presidential Democracies Compared. European Journal of Political Research 48(5): 665–693.

56.

Sheng

Shing-Yuan

(2003) The Influence of Legislative and Administrative Institutions in the Legislative Process: A comparison of unified and divided government. Taiwanese Journal of Political Science 7(2): 51–105.

57.

Shugart

Matthew Soberg

Carey

John Michael

(1992) Presidents and Assemblies: Constitutional Design and Electoral Dynamics. Cambridge: Cambridge University Press.

58.

Skach

Cindy

(2007) The ‘Newest’ Separation of Powers: Semi-presidentialism. International Journal of Constitutional Law 5(1): 93–121.

59.

Solt

Frederick

(2009) Standardizing the World Income Inequality Database. Social Science Quarterly 90(2): 231–242.

60.

Steenbergen

Marco

Jones

Bradford

(2002) Modeling Multilevel Data Structures. American Journal of Political Science 46(1): 218–237.

61.

Steinmo

Sven

(1994) Rethinking American Exceptionalism: Culture or institutions? In Dodd

Lawrence C

Jilsson

Calvin C

(eds) The Dynamics of American Politics: Approaches and Interpretations. Boulder, CO:Westview Press.

62.

Stepan

Alfred

Skach

Cindy

(1993) Constitutional Frameworks and Democratic Consolidation: Parliamentarianism versus presidentialism. World Politics 46(1): 1–22.

63.

Stepan

Alfred

Suleiman

Ezra

(2001) The French Fifth Republic: A model for import? Reflections on Poland and Brazil. In Stepan

Alfred

(ed) Arguing Comparative Politics. Oxford: Oxford University Press.

64.

Strøm

Kaare

Mueller

Wolfgang

Bergman

Torbjörn

Nyblade

Benjamin

(2003) Dimensions of Citizen Control. In Strøm

Kaare

Müller

Wolfgang C

Bergman

Torbjörn

(eds) Delegation and Accountability in Parliamentary Democracies. Oxford/New York: Oxford University Press.

65.

Tsai

Jung-Hsiang

(2008) Sub-Types of Semi-Presidentialism and Political Deadlock. French Politics 6(1): 63–84.

66.

Verba

Sidney

Schlozman

Kay Lehman

Brady

Henry

(1995) Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.

67.

Chung-Li

(2009) Semi-Presidentialism and Divided Government in Taiwan: Public perceptions of government performance. Issues and Studies 45(2): 1–34.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.38 MB