Presidential Power, Party System and Cabinet Instability in Semi

Presidential Power, Party System and Cabinet Instability in
Semi-presidentialism
Tsai, Jung-hsiang
What distinguishes semi-presidentialism from presidentialism and from
parliamentarism is dual executive. The president can unilaterally appoint the prime
minister or can nominate the prime minister but has to get the investiture of the
parliament. Not only can the president maneuver the formation of cabinet but also has
authority on the dismissal of the cabinet and the prime minister. The president
possesses some formal power or informal authority to dismiss the prime minister,
especially when the prime minister is from the same party (Samuels and Shugart,
2010). However, the parliament also has the censure power to sack the prime minister
in semi-presidentialism. Comparatively speaking, we can envision that semipresidentialism entails more cabinet instability than presidentialism or than
parliamentarism. Cabinet instability can indicate policy failure, policy discontinuity,
or intraexecutive conflict in semi-presidentialism. When some landmark policies fail,
the cabinet can be held accountable and can be dismissed afterwards. Frequent
turnovers of cabinets would also make some policies lagged or even not protracted.
Furthermore, cabinet instability could make it difficult for governments to adopt and
implement new policy programs and portfolio volatility could make it difficult for
cabinet ministers to obtain relevant information during policy formation and
implementation (Huber, 1998). Simply put, cabinet instability produces damaging
effects on the continuity and effectiveness of policy making in semi-presidentialism.
Fierce struggles between the president and prime minister can also spark the
reformation of the cabinet. Cabinet instability turns out to be a deterrent to democratic
governance and even leads to democratic breakdown in some semi-presidential
countries (Bernhard, 2005; Powell, 1982; Skach, 2005). Why some cabinets cannot
finish their terms and have to be terminated prematurely as some cabinets endure until
the next election approaches in semi-presidentialism is the major puzzle in this paper.
This paper seeks to elucidate the circumstances under which cabinets break down,
what structural attributes cause more cabinet instability in semi-presidentialism, and
how the fluctuation of cabinets affects the operation of semi-presidentialism. By
keeping close track of these pathways systematically, it can increase our
understanding of cabinet turnovers and their effects on governance of semipresidentialism.
A number of factors can bring out cabinet instability in semi-presidentialism.
One of them originates from the institutional features of semi-presidentialism. When
the president has the authority to dismiss the prime minister, it can cause more cabinet
instability than the president without that prerogative, all else equal. In presidentparliamentary systems, one subtype of semi-presidentialism, the president is
constitutionally granted the power to nominate and dismiss the prime minister
(Shugart, 2005; Shugart and Carey, 1992). Coupling with these powers, the president
in president-parliamentary systems can pull strings to the formation and termination
of cabinets. In premier-presidential systems, the other subtype of semipresidentialism, the president does not hold the formal power to terminate the cabinet
but still can wield his informal power such as partisan influence to sack the prime
minister. For example, in France, of the 18 changes on prime minister since the 5 th
Republic, 12 times of the dismissal of prime minister were due to presidential partisan
influence (Samuels and Shugart, 2010: 104). The president in France can wield his
informal power to the fullest as the prime minister is from the president’s party. If the
prime minister is from the opposing party such as the scenario in the phase of
1
cohabitation, presidential partisan influence on the deselection of prime minister has
completely waned or even disappeared. Taken together, the president in presidentparliamentary systems possesses formal and informal powers on cabinet survival or
dismissal while the president in premier-presidential systems only holds informal
powers on that matter.
Past research suggests that intraexecutive conflicts can increase the possibility of
cabinet turnovers in semi-presidentialism (Sedelius, 2006: 227-34; Sedelius and
Ekman, 2010). This approach has tended to focus on conflictual politics but to
downplay the importance of structural powers of the president and cabinet. Whether
the president with some structural powers allows him to have open conflict with the
prime minister coordinating the cabinet or the prime minister and his party that have
an advantage over the other coalition party or the president’s party in the cabinet are
more likely to cross swords with the president. This paper will demonstrate that the
occurrence of cabinet instability relies significantly on the balance of structural
powers between the president and parties in the cabinet or in the parliament.
There are two types of terminations of coalition cabinets, namely, technical
terminations and discretionary terminations.1 This paper mainly focuses on
discretionary terminations, not technical terminations. Furthermore, we are more
concerned with the termination of cabinets which is involved with the president in
semi-presidentialism. In order to follow the logic of comparison, we calculate the
number of discretionary terminations which occurs in any presidential term. Different
semi-presidential countries may have different presidential terms. In practice, we
divide the number of discretionary terminations by the years of a presidential term
and we can get an average of cabinet turnovers per year. 2 We endeavor to do this
throughout the paper by first, reviewing the literature of cabinet instability in
parliamentarism and in presidentialism; second, highlighting our theoretical reasoning
about how the variation of cabinet instability is related to the balance of the president
and other parties and the circumstances under which we are likely to see imbalance of
structural powers leads to the turnover of cabinets and then propose our hypotheses to
be tested in this paper; third, describing how we collect the data and what kind of
methods we employ to analyze the data; and fourth, concluding our preliminary
findings.
The Literature of Cabinet Instability
Semi-presidentialism is characterized by the definition which a president who is
1 Technical terminations are beyond the control of political parties in the coalition such as regular
elections, the election of a new president that requires the cabinet to resign, or an accident of the prime
minister; discretionary terminations are deliberately brought about by political parties in the coalition
such as general elections called before the end of the regular term, voluntary enlargements of the
cabinet to include previously unrepresented parties, prime ministerial resignations caused by votes of
no confidence in parliament, voluntary resignations by the prime minister for other political reasons,
cabinet resignations caused by conflicts with cabinet parties, and terminations caused by conflict
between coalition parties for policy and/or personal reasons (Damgaard, 2008: 303-4).
2 Lee (2012) explores the effects of presidential power and parliamentary composition on the duration
of prime minister in semi-presidential democracies. The dependent variable in her study is how many
days the prime minister holds the position. Her major findings are that the president with the formal
power to dismiss the prime minister can pose more risks to the duration of prime minister than the
president without the formal power to dismiss the prime minister and the more legislative seats the
president party holds in the coalition cabinet, the less possible the prime minister is dismissed.
However, this study focuses on the frequency of presidential termination on prime minister instead of
focusing on the time length of the prime ministership. In so doing, we can exclude technical
terminations and we can treat each presidential term as a case.
2
popularly elected fixed term and a prime minister and cabinet who are responsible to
parliament (Elgie, 1999:13). It combines some features of presidentialism and some
features of parliamentarism. For research purposes, we can draw on the theory of
government termination in parliamentarism and the literature on presidential cabinets
in presidentialism to develop an analytical framework about cabinet instability in
semi-presidentialism (Tsai, 2011:4-5).
The most salient feature of parliamentarism is the monopoly of power. When one
party controls a majority of seats after parliamentary elections, its party leader
assumes the position of prime minister. The ruling party simultaneously holds the
control of the executive and legislative branches. If no party controls a majority of
legislative seats, two or more parties would unite for a coalition cabinet to share
political power. What factors affect the termination of coalition cabinets has driven
many scholars to explore. One study focuses on the hope factor in the future. Parties
terminate cabinets when they anticipate more electoral gains and more policy gains in
the new elections (Grofman and van Roozendaal, 1994). It demonstrates that rational
calculation and strategic interaction of the parties in the coalition are the major force
at work to lead to the turndown of cabinets. Another study claims that the more the
number of cabinet parties and the greater ideological diversity it is in the coalition, the
more likely cabinet instability is to appear (Warwick, 1992). On the same token,
conflictual terminations occur more when bargaining power in the cabinet is dispersed
and when the party with the greatest bargaining power is not represented in the
cabinet (Damgaard, 2008). To synthesize, these studies lay stress on how the
structural attributes to cause cabinet instability in the coalition government in
parliamentarism.3 We will assess cabinet instability in semi-presidentialism through
the prism of structural power in the coalition and explain why some cabinets undergo
a collapse or a reformation and why some cabinets do not.
Research on presidential countries in Latin America finds that cabinet instability
is more likely when parties find it costly to participate in the government and the
president with strong unilateral institutional powers tends to produce more cabinet
instability than the president without those powers (Martinex-Gallardo, 2012).
Accordingly, stronger presidential authority is related to higher cabinet instability
since presidents with strong decree powers have fewer incentives to compromise
(Martinex-Gallardo, 2012: 78). However, apart from presidential formal powers,
strong or weak presidential authority is largely contingent on the types of cabinet and
the seats of legislative control. In presidential regimes, the president has full
discretion to form the cabinet and to appoint the portfolio. If the president’s party
holds a legislative majority of seats in the parliament, the president would form a
single-party cabinet filled with co-partisans and loyalties. If the president’s party lacks
a legislative majority, the president can form a minority cabinet or a coalition cabinet.
The challenge of a minority cabinet is that it has to extend legislative support from
other parties in the parliament in order to pass any bill. Instead of governing alone, the
president with a minority of legislative support in multiparty presidential countries of
Latin America would form a multiparty cabinet which includes the president’s party
and other coalition parties in order to get bills or proposals smoothly passed in the
congress (Amorim Neto, 2002). Among coalition cabinets in presidentialism, surplus
3 Some study focuses more on how exogenous events affect the turnover of cabinets than on structural
factors (Browne, Frendreis, and Gleiber, 1984). Methodologically, the study of cabinet instability in
parliamentary systems has two research traditions. One is the empirical approach which uses eventhistory model and the other is the game theory which emphasizes the importance of strategic
interaction (Laver, 2003).
3
coalitions are less durable since any one of the government parties can defect without
jeopardizing the government’s majority so that presidents are less willing to strike a
new bargain with their coalition partners; on the contrary, minimum winning
coalitions (MWC) are more durable since one party would leave the government
without a parliamentary majority so that presidents are more willing to strike a new
bargain with their coalition partners (Martinex-Gallardo, 2012: 70).
In presidentialism, the government cannot be turned down by Congress and the
turndown of cabinets does not affect the survival of presidents (Stepan and Skach,
1993). The benefit of maintaining a legislative majority by cementing a coalition is
more likely to pass bills in the congress. In the scenario of minimum winning
coalitions, the president has to compare the cost of persuading a possible defecting
party to stay in the coalition with the cost of losing a legislative majority after that
party chooses to leave. If a coalition party chooses to leave the cabinet, the president
still has two options. One is to maintain a minority cabinet but to build ad hoc
legislative coalitions on a case-by-case account. The other is to recruit another party
to form a different coalition with a new legislative majority. Simply put, for presidents
in presidentialism, any party defecting from the coalition affects the reshuffle of
cabinets or the change of legislative support, not the survival of the president. This is
to say that presidents in presidentialism can tolerate more cabinet instability than the
prime minister in parliamentarism, all else being equal. In terms of our concern here,
semi-presidentialism has dual leaders: the president and prime minister. The operation
of cabinets can possibly strike a middle course between that in presidentialism and
that in parliamentarism. Cabinet instability indicates the termination of the prime
minister’s tenure but it cannot affect the survival of the president.
Cabinet Formation and Termination in Semi-presidentialism
The underlying process that leads to the formation or turnover of cabinet in semipresidentialism has different configurations. Some presidents in semi-presidentialism,
especially in president-parliamentary systems, can nominate the prime minister
without the consent of parliament and can also dismiss the cabinet or sack the prime
minister at will (Shugart, 2005; Shugart and Carey, 1992). Basically, the president is
the formateur who can call the shots on the formation and termination of the cabinet.
However, in some president-parliamentary countries, the nomination of prime
minister has to get the investiture of the parliament such as Russia. The president may
nominate the prime minister but the final outcome is a seesaw fight between the
president and parliament. As the president chooses to discharge the prime minister, he
has to consider whether a majority in the parliament can accept the new nominee or
not. If the nominee of prime minister fails to pass the threshold of investiture, the
president has to nominate another candidate. In all president-parliamentary countries,
not only can the president dismiss the cabinet, but also the parliament can cast a vote
of no-confidence to censure the cabinet. It is a balance of terrors between the
president and opposing parliament on the survival or breakdown of cabinets. When
the parliament threatens to propose a motion of censure, the president can
preemptively terminate the current cabinet and replace a new one. If the parliament
passes the motion of censure, the president can respond to dissolve the parliament and
stage a new parliamentary election. In sum, the mechanism of cabinet formation and
deselection in semi-presidentialism is mutual jeopardy between the president and
parliament (Shugart, 1993).
One recent study on East and West European cabinets finds that the president
power to dismiss a government has no effect on the dissolution hazard that
4
government faces (Schleiter and Morgan-Jones, 2009: 507). Why Schleiter and
Morgan-Jones get this conclusion is something to do with their dataset. Schleiter and
Morgan-Jones only include two president-parliamentary systems in the total 18 semipresidential cases. 16 countries are premier-presidential systems. The president in
premier-presidential systems has no formal power to dismiss the government so that it
is likely to reach the conclusion that the presidential power to dismiss the government
has no effect on the dissolution hazard that governments face. If we enact the method
of controlled comparison and compare all the premier-presidential systems with all
the president-parliamentary systems in terms of dissolution hazard, we may get a
different picture as shown in H1. In addition, all else being equal, within presidentparliamentary systems, the countries with a multiparty system have more hazard rates
on cabinet instability than the countries with a two-party system, as shown in H2.
H1: President-parliamentary systems can have more hazard rates on cabinet
instability than premier-presidential systems.
H2: President-parliamentary countries with a multiparty system have more hazard
rates on cabinet instability than president-parliamentary countries with a two-party
system.
Presidential constitutional powers are highly correlated with the formation of
cabinets. Strong presidential authority in the constitution can increase the number of
nonpartisan ministers nominated by the president and pull in the direction of
president-controlled cabinets rather than the direction of assembly-controlled cabinets
(Schleiter and Morgan-Jones, 2010). Another study on seven European countries
(Austria, Iceland, France (since 1958), Germany(since 1949), Iceland, Italy and
Portugal (since 1976)) finds that the president is more likely to nominate a co-partisan
member as the prime minister and the coalition cabinet is more likely to form if it
includes the presidential party (Kang, 2009). Taken together, the power of
appointment can give the president a political edge and keep the prime minister and
his or her cabinet more to his liking in office. For presidents in some semi-presidential
countries, not only do they hold the power of nomination but also grasp the power to
sack the prime minister. The president with these two powers can be preponderant on
the formation and termination of cabinet if he also enjoys the support of a
parliamentary majority. When the president’s party only holds minority seats in the
cabinet or holds no seats in the cabinet (such as cohabitation), the president’s
constitutional authority can be largely dwindling. It means that the president even
with strong constitutional authority still has to face the variation of political dynamics
in the parliament. When the president opts to dismiss the cabinet, he has to negotiate
with coalition parties in the cabinet. Unlike single-party cabinets, the termination of
multiparty cabinets in semi-presidentialism involves strategic interactions among the
president, the president’s party, and other coalition parties. The enactment of
presidential powers has to be considered with the party composition of cabinets or
with the variation of party systems in the parliament. Looking at the interaction of
presidential structural powers and political dynamics in the cabinet or parliament adds
an important strand that has been mostly absent from the literature on semipresidentialism. The proposition that is derived from our reasoning is that:
H3: The president with the power to dismiss the government poses more hazard hates
on cabinet termination in semi-presidentialism.
5
Conventional wisdom claims that semi-presidentialism cannot work very well
with a highly fragmented party system or a polarized multiparty system (Linz, 1997).
Empirically, moderate party system fragmentation (the effective number of parties:
2.47-3.06) and high level of party system fragmentation (the effective number of
parties: 3.07-4.37) can increase the hazard of democratic collapse in semi-presidential
countries (Elgie and Schleiter, 2011, 55-56). However, the fragmentation of party
system cannot cause democratic collapse directly. An intervening variable such as
cabinet instability works in between the fragmentation of party system and democratic
collapse in semi-presidentialism.4 Cabinet instability is caused by the fragmentation of
party system5 and cabinet instability is a cause of democratic collapse. This paper
underlines the interaction between the fragmentation of party system and cabinet
instability. The hypothesis below is derived from the discussion.
H4: The higher the number of effective parties in the parliament, the more possible
the cabinet is to collapse or vice versa.
Control Variables
In this paper we also control some related variables which can affect cabinet
instability. First, exogenous factors such as economic crises can be shocks to the
duration of cabinet. When the economy is in recession, cabinet turnovers can be more
frequent than the economy is in a good shape since the cabinet can be blamed
politically. Specifically, two economic indicators such as unemployment rate and
inflation affect the likelihood of premature or conflictual cabinet termination and high
unemployment and low inflation would make cabinets more fragile (Damgaard, 2008:
321). Seen in this light, we adopt the economic indicator (GDP per capita) indicators
in the first year of the president’s tenure. Countries with more heterogeneous ethnic
groups are more likely to be polarized than countries with homogenous ethnic groups.
The aggregate data of ethnic fractionalization developed by Alesina et al (2002) are a
useful indicator as controlling the effects of other variables on cabinet instability. We
also control the staggering feature of presidential elections and prime ministerial
elections. If presidential elections and prime ministerial elections are concurrent, then
we would expect less cabinet instability. If presidential elections and prime ministerial
elections are staggering, then more cabinet instability can occur, other things being
equal.
Dependent Variable
The number of semi-presidential countries has been rising to keep pace with the
number of presidential countries and parliamentary countries. The origin of semipresidentialism comes from combining some feature of presidentialism and of
parliamentarism such as a president who is popularly elected fixed term and a prime
minister and cabinet who are responsible to parliament (Elgie, 1999: 13). However,
semi-presidentialism evolves into a new species not a mixed type since its operation
can be involved in a triangular relationship among the president, the prime minister,
4 Another possible intervening variable is legislative paralysis or immobolism which can make
democracy less responsive to political crises.
5 According to a recent study, ruling parties’ hazard rates under the first-past-the-post systems are
initially lower than those under proportional representation rules and ruling parties’ hazard rates under
two-party systems are initially lower than those under multiparty systems (Nishikawa, 2012).
6
and the parliament which differs from the other two constitutional types. As more
countries have chosen semi-presidentialism as their constitutional type, they have
provided rich data for testing the feasibility of semi-presidentialism.
In past work on semi-presidentialism, there have been three different lines of
research. The first line analyzes crucial cases, such as France, Portugal, and the
Weimar Republic, in depth (Pasquino, 1997; Skach 2005; Suleiman 1994). The
second line constructs different subtypes to explicate the stability or instability of
semi-presidential democracies (Shugart 2005; Tsai 2008; Wu, 2000; Shugart and
Carey, 1992). The third line uses the large-N datasets to detect or verify what possible
factors affect the operation of semi-presidentialism (Elgie, 2011; Kang, 2009; Amorim
Neto and Strøm, 2006; Cheibub and Chemykh, 2008; Elgie and McMenamin, 2008;
Elgie and Schleiter, 2011; Schleiter and Morgan-Jones, 2009; 2010). Each of these
approaches has its uses and limitations. This paper will draw on data from all semipresidential democracies and gauge structural factors to cause the frequency and
hazard of cabinet instability by controlling other relevant variables.
We are mostly concerned with one dependent variable in this study. Our
dependent variable is how many times the president dismisses the cabinet during
his/her term. The president can serve for one term or two terms or more. We take an
average of the number of discretionary terminations per year.
Data
Our data cover all semi-presidential democratic countries observed from 1945 or from
the adoption of a semi-presidential constitution until 2012. A couple of the datasets we
can employ for our purpose. We only choose partial free and free semi-presidential
countries in the Freedom House dataset. For party affiliation of the president and
prime
minister,
we
use
the
website
of
worldstatesmen
(http://www.worldstatesmen.org/). On presidential power, we follow previous studies
such as Shugart and Carey (1992) and Siaroff (2003). Regarding the information of
parliamentary elections and the seats of the president party we rely on the Polity IV
project: political regime characteristics and transitions, 1800-2010, the database of
Political Institutions, and Adam Carr’s election archive (http://psephos.adam-carr.net/)
etc. As for coding the dependent variable, cabinet instability, this paper utilizes the
website of worldstatesmen which contains the information of presidential terms and
the terms of prime ministers in all semi-presidenitalism. We exclude the prime
ministers who are produced by parliamentary elections and only take discretionary
terminations into account. Totally, we have 159 cases (presidential terms).
Analysis
We compare the sample means of the frequency of the dismissal to the prime minister
in each presidential term for two different subtypes in semi-presidentialism. The H1
hypothesis suggests that when we divide the sample on the basis of different
constitutional structures in semi-presidential countries and compare mean values of
the frequency of the dismissal to the prime minister in each presidential term, the
mean of president-parliamentary systems will be higher than the mean of premierpresidential systems. The null hypothesis asserts that no group differences exist in the
population of semi-presidential countries. The Independent-Samples T test allows us
to test the H1 and the null hypothesis and to determine whether the observed
differences between president-parliamentary systems and premier-presidential
systems are too large to have occurred by random sampling error.
From Table 1, we can see that president-parliamentary systems, on average, rated
the frequency of dismissal at .3457, whereas premier-presidential systems, on
7
average, had a lower mean, .2200. It demonstrates that the H1 hypothesis not the null
hypothesis can hold up here. The difference between these two sample means is .3457
minus .2200, or .12568. Please see the column of Mean Difference of the
Independent-Samples Test, which is shown as Table 2. First, judged by the
Independent-Samples Test table, we try to compare the merits between the H1and the
null hypothesis, concluding that the mean of president-parliamentary systems is
higher than the mean of premier-presidential systems on the dismissal rate of the
prime minister. Second, through “Levene’s Test for Equality of Variances, if the value
is greater than or equal to .05, then we have to use the “Equal variance assumed” row
or if the value is lower than .05, we choose to use the “Equal variance not assumed”
row. Since the “Sig.” value for the frequency of dismissal is lower than .05, we use
the second row of numbers to gauge the mean differences between presidentparliamentary systems and premier-presidential systems. Third, we test our hypothesis
by the 90 percent confidence interval of the mean difference to apply the .05
threshold. Since the direction of the mean difference is consistent with the H1 and the
confidence interval does not include 0, the mean difference is statistically significant
at the .05 level. Fourth, we can get a probability, or P-value, by dividing the “Sig. (2tailed) value by 2. The value in the “Sig. (2-tailed)” column is .011. .We divide .011
by 2, which can get .005. If the null hypothesis is right, then random sampling error
could produce a mean difference of .12568 only by .005 percent of chance. This result
bears out the H1 and allow us to reject the null hypothesis.
Table 1. Group Statistics
Std.
Subtype
s
Deviatio
N
Mean
n
Std. Error Mean
Frequen 1.00
81
.3457
.33666
.03741
cy
78
.2200
.27934
.03163
.00
Table 2. Independent Samples Test
Levene's Test for Equality of
Variances
t-test for Equality of Means
90% Confidence Interval of
F
Frequency
Equal variances
Sig.
7.049
t
.009
df
Sig. (2-
Mean
Std. Error
tailed)
Difference
Difference
the Difference
Lower
Upper
2.557
157
.012
.12568
.04916
.04434
.20702
2.566
153.672
.011
.12568
.04899
.04462
.20674
assumed
Equal variances not
assumed
8
Table 3. Model Summary
Model
R
1
.351
Adjusted R
Std. Error of the
Square
Estimate
R Square
.123
a
.100
.29903
a. Predictors: (Constant), Economic, ENP, Categorical, Polarization
Table 4. ANOVAb
Model
1
Sum of Squares
Regression
df
Mean Square
1.932
4
.483
Residual
13.771
154
.089
Total
15.703
158
F
Sig.
5.403
.000a
t
Sig.
a. Predictors: (Constant), Economic, ENP, Categorical, Polarization
b. Dependent Variable: Frequency
Table 5. Coefficientsa
Standardized
Unstandardized Coefficients
Model
1
B
Std. Error
Coefficients
Beta
(Constant)
.043
.111
Categorical
.140
.053
.223
2.634
.009**
ENP
.036
.012
.229
2.958
.004**
-.077
.123
-.063
-.628
.531
-1.138E-5
.000
-.210
-2.271
.025**
Polarization
Economic
.391
.697
a. Dependent Variable: Frequency; *=p < 0.1; **=p < .05; ***=p <.01.
After reporting the Independent-Samples test, we execute a multiple regression
which can estimate the partial effect of an independent variable while controlling for
the effects of other independent variable. We use frequency as the dependent variable
and entering categorical indicator (subtypes of semi-presidentialism), effective
number of parties (ENP), polarization, and economic indicators. Categorical indicator
and ENP are explanatory variables while polarization and economic indicator are
control variables. The multiple regression equation is as follows:
Frequency=
0.43+.140*(subtypes of semi-presidentialism)+.036(ENP)-.077(polarization)-1.138E5(economic).
9
From the regression output (Table 3, Table 4, and Table 5), we can conclude that
president-parliamentary regimes can produce more cabinet instability in terms of
dismissal frequency on the prime minister and the higher the number of effective
parties in the parliament, the more possible the cabinet is to collapse while controlling
two variables (polarization and economic indicator). The subtype variable was
significantly predictive of the frequency of cabinet instability when the other variables
were statistically controlled: t (155) =2.634, p=0.009. The ENP variable was also
significantly predictive of the frequency of cabinet instability when the other variables
were statistically controlled: t (155) =2.958, p=0.004. Regarding H2, we put an
integrated variable which combines different subtypes with different effective
numbers of parties. We also put another control variable which is staggering elections.
Close examination of the data for a sample of N=159 cases indicated that the
correlation between our integrated variable and other variables did not indicate any
multicollinearity. The output was shown as Table 7, Table 8, and Table 9. The
integrated variable was significantly predictive of the frequency of cabinet instability
when the other variables were statistically controlled: t (155) =3.396, p=0.001. In
other words, president-parliamentary countries with a multiparty system have more
hazard rates on cabinet instability than president-parliamentary countries with a twoparty system, premier-presidential system with a multiparty system, or premierpresidential system with a two-party system.
Table 6. Correlations
Categorical*EN
Economic
Economic
Pearson Correlation
P
1
-.576**
.492
.003
.000
159
159
159
159
-.055
1
.089
.074
.263
.352
N
Staggering
Polarization
Pearson Correlation
Sig. (2-tailed)
.492
N
159
.233
159
159
159
**
.089
1
-.411**
Sig. (2-tailed)
.003
.263
N
159
159
Pearson Correlation
Pearson Correlation
.233
-.576
.000
159
**
.074
Sig. (2-tailed)
.000
.352
.000
N
159
159
159
**. Correlation is significant at the 0.01 level (2-tailed).
10
-.055
Polarization
**
Sig. (2-tailed)
Categorical*ENP
Staggering
-.411
**
159
1
159
Table 7. Model Summary
Model
R
2
Adjusted R
Std. Error of the
Square
Estimate
R Square
.363
a
.132
.109
.29750
a. Predictors: (Constant), Staggering, Categorical*ENP, Economic,
Polarization
Table 8. ANOVAb
Sum of
Model
2
Squares
df
Regression
Mean Square
F
Sig.
2.073
4
.518
Residual
13.630
154
.089
Total
15.703
158
5.857
.000a
Table 9. Coefficientsa
Standardized
Unstandardized Coefficients
Model
1
B
(Constant)
.130
.101
Polarization
.033
.121
Categorical*ENP
.025
Economic
Staggering
a. Dependent Variable: Frequency
11
Std. Error
Coefficients
Beta
t
Sig.
1.285
.201
.027
.273
.785
.007
.258
3.396
.001
-1.145E-5
.000
-.211
-2.298
.023
.104
.068
.127
1.531
.128
Discussion
The results of parliamentary elections significantly affect the type of cabinet. If one
party grasps a majority of seats in the parliament, most of the time it can form a
single-party cabinet. In this scenario, the president’s margin of maneuver on the
formation of cabinet would increase extensively. However, when there is no party
gaining a majority, the president has to build a coalition cabinet to seal an alliance
with other parties in parliament. In the coalition cabinet, if the president’s party
controls relatively more seats than any other coalition party, the president and his
party can still take the initiative on policy making even though they cannot be
dominant on all counts. If the president’s party is a minor party in the cabinet and the
prime minister is from a major party in the coalition, then presidential partisan
influence has been largely weakened in the process of decision-making within the
coalition. Under this scenario, the president has to maintain a delicate balance with
the coalition party or parties in order to enforce his preferred policies. Whether the
president with dismissal power of prime minister in this scenario tends to sack the
prime minister he mostly disagrees with or the president tends to maintain the stability
of cabinet since his party strength in the cabinet is inferior to the other coalition party
remains to be explored in the future.
Endogenous changes of coalitions can also cause cabinet instability. When a
coalition party chooses to exit and then the cabinet lacks a legislative majority to
command, the president has to form another cabinet which can solidify enough seats
in parliament in order to pass any bill. Another similar situation is that a coalition
party threatens to exit and asks for higher utility such as more ministerial posts or
more spending and it would squeeze the president’s party out and curtail its spending
on targeted voters. Since maintaining such a coalition is costly, the president would
dismiss the cabinet rather than being blackmailed by the coalition party or parties.
Taken together, when the coalition is hard to cement or when it is hard for the
president to make concessions to unpalatable coalition proposals, the president would
press the button to recall the cabinet and find a new workable cabinet.
The pioneer study on the subtypes of semi-presidentialism conducted by Shugart
and Carey (1992) claims that president-parliamentary systems work worse than
premier-presidential systems. Even they warn to stay away from presidentparliamentary systems (Shugart and Carey, 1992: 287). Elgie (2011) provides an
empirical support that president-parliamentary systems are more perilous to
democratic survival than premier-presidential systems (see also Elgie and Schleiter
(2011)). Elgie (2011:33-4) argues that in president-parliamentary systems, the
president and the parliament can jockey for the composition of a new cabinet and
cause a zero-sum game against each other. Our study here echoes some partial
findings of the conventional wisdom on the performance of semi-presidentialism.
President-parliamentary systems can cause more cabinet instability than premierpresidential systems. However, president-parliamentary system only with a multiparty system can be a difficult combination. If president-parliamentary systems
coupled with low levels of multi-partyism or two-party systems can conversely entail
less cabinet instability. Frequent cabinet instability can pose more threats on
democratic governance of semi-presidentialism. The causal mechanism of how
president-parliamentary systems and multi-party systems go together to make
democracy more vulnerable is the agenda for future research.
12
Reference
Alesina, A, A. Devlesschauwer, W. Easerly, S. Kurlat and R. Wacziarg. 2002.
Fractionalisation. NBER Working Paper Series Working Paper 9411.
Amorim Neto, Octavio. 2002. “Presidential Cabinets, Electoral Cycles, and Coalition
Discipline in Brazil.” in Scott Morgenstern and Benito Nacif eds., Legislative
Politics in Latin America. Cambridge: Cambridge University Press.
Amorim Neto, Octavio. 2006. “The Presidential Calculus: Executive Policy Making
and Cabinet Formation in the Americas,” Comparative Political Studies, Vol. 39.
N0.4: 415-40.
Bernhard, Michael, 2005. Insitutions and the Fate of Democracy. Pittsburgh:
University of Pittsburgh Press.
Browne, Eric C, John P. Frendreis and Dennis W. Gleiber, 1984. “An Events
Approach to the Problem of Cabinet Stability,” Comparative Political Studies
17(2): 167-97.
Caplow, Theodore. 1956. “A Theory of Coalitions in the Triad,” American
Sociological Review 21(4): 489-93.
Cameron, Charles. 2000. Veto Bargaining: Presidents and the Politics of Negative
Power. Cambridge: Cambridge University Press.
Carey, John and Matthew Shugart. 1998. Executive Decree Authority. Cambridge:
Cambridge University Press.
Cheibub, Jose and S. Chernykh. 2008. “Constitutions and Democratic Performance in
Semi-presidential Democracies,” Japanese Journal of Political Science 9(3): 269303.
Cox, Gary W and Scott Morgenstern. 2002. “Latin America’s Reactive Assemblies
and Proactive Presidents,” Comparative Politics 33(2): 171-89.
Damgaard, Eric. 2008. “Cabinet Termination,” in Kaare Strom, Wolfgang C. Muller
and Torbjorn Bergman. eds., Cabinets and Coalition Bargaining: The Democratic
Life Style in Western Europe, New York: Oxford University Press, pp. 301-26.
Elgie, Robert. 1999. “The Politics in Semi-presidentialism,” In Robert Elgie, ed.,
Semi-presidentialism in Europe. New York: Oxford University Press, pp. 1-21.
Elgie, Robert and Iain McMenamin. 2008. ‘Semi-presidentialism and Democratic
Performance,” Japanese Journal of Political Science. Vol. 9, No.3: 323-40.
Elgie, Robert and Petra Schleiter, 2011. “Variation in the Durability of Semipresidential Democracies,” in Robert Elgie ets, Semi-presidentialism and
Democracy. London: Palgrave Macmillan Press, 42-60.
Gamson, William A. 1961. “A Theory of Coalition Formation,” American
Sociological Review 26(3): 373-382.
Grofman, Bernard, and Peter van Roozendaal. 1994. “Toward a Theoretical
Explanation of Premature Cabinet Termination.” European Journal of Political
Research, 26: 155-70.
Huber, John. 1998. “How Does Cabinet Instability Affect Political Performance?
Portfolio Volatility and Health Care Cost Containment in Parliamentary
Democracies,” American Political Science Review, 92(3): 27-48.
Kang, Shin-koo. 2009. “The Influence of Presidential Heads of State on Government
Formation in European Democracies: Empirical Evidence,” European Journal
of Political Research, 48: 543-72.
Laver, Michael. 2003. “Government Termination,” Annual Review of Political
Science, 6: 23-40.
13
Lee, Yi-fang. 2012. The Duration of Prime Minister in Semi-Presidentialism : The
Effect of Presidential Power and Composition of Parliament. Thesis for political
science department, National Taiwan University, Taipei.
Linz, Juan. 1997. “Introduction: Some Thoughts on Presidentialism in Postcommunist
Europe.” in Ray Taras ed., Postcommunist Presidents. Cambridge: Cambridge
University Press,1-14.
Lupia Arthur and Kaare Strom. 1995. “Coalition Termination and the Strategic Timing
of Parliamentary Elections,” American Political Science Review,Vol. 89, No. 3:
648-65.
Martin, Lanny W. and George Vanberg. 2011. Parliaments and Coalitions: The Role
of Legislative Institutions in Multiparty Governance. Oxford: Oxford University
Press.
Martinez-Gallardo. 2012. “Out of the Cabinet: What Drives Defections From the
Government in Presidential Systems?,” Comparative Political Studies 45(1):
62-90.
Morgan-Jones, Edward and Petra Schleiter. 2004. “Governmental Change in a
President-Parliamentary Regime: The Case of Russia 1994-2003,” Post-Soviet
Affairs, 20(2): 132-173.
Nishikawa, Misa. 2012. “Electoral and Party System Effects on Ruling Party
Durability,” Party Politics, 18(5): 633-52.
Pasquine, Paul.1997. “Semi-presidentialism: A Political Model at Work,” European
Journal of Political Research 31: 128-37.
Protsyk, Oleh. 2005. “ Politics of Intraexecutive Conflict in Semi-presidential
Regimes: Constitutional Norms and Cabinet Formation Outcomes,” East
European Politics and Societies 19(2): 135-60.
Powell, Bingham Jr. 1982. Contemporary Democracies: Participation, Stability and
Violence. Cambridge, Mass.: Harvard University Press.
Remington, Thomas, Steven Smith, and Moshe Haspel. 1998. “Decrees, Laws, and
Inter-branch Relations in the Russian Federation,” Post-Soviet Affairs, 14(4): 287322.
Samuels, David and Matthew S. Shugart 2010. Presidents, Parties, and Prime
Ministers: How Separation of Powers Affects Party Organization and Behavior.
New York: Cambridge University Press.
Schleiter, Petra and Edward Morgan-Jones. 2009. “Constitutional Power and
Competing Risks: Monarchs, Presidents, Prime Ministers, and the Termination of
East and West European Cabinets,” American Political Science Review 103(3):
496-512.
Schleiter, Petra and Morgan-Jones, Edward. 2010. “Who’s in Charge? Presidents,
Assemblies, and the Political Control of Semi-presidential Cabinets,”
Comparative Political Studies, 43(11): 1415-1441.
Shugart, Matthew S. 1993. “Of Presidents and Parliaments,” East European
Constitutional Review 2(1): 30-2.
Shugart, Matthew S. 2005. “Semi-presidential Systems: Dual Executive and Mixed
Authority Patterns,” French Politics, Vol. 3, No.3: 323-51.
Shugart, Matthew S. and John M. Carey. 1992. Presidents and Assemblies:
Constitutional Design and Electoral Dynamics. Cambridge: Cambridge
University Press.
Skach, Cindy. 2005. Borrowing Constitutional Designs: Constitutional Law in
Weimar Germany and the French Fifth Republic. New Jersey: Princeton
14
University Press.
Stepan, Alfred, and Cindy Skach, 1999. “Constitutional Frameworks and Democratic
Consolidation. Parliamentarism and Presidentialism,” World Politics 46(1): 122.
Sulieman, Ezra A. 1994. “Presidentialism and Political Stability in France.” In Juan
Linz and Arturo Valenzuela, The Failure of Presidential Democracy. Baltimore:
Johns Hopkins University Press, 137-62.
Tsai, Jung-hsiang. 2008. “Subtypes of Semi-presidentialism and Political Deadlock,”
French Politics, 6(1): 63-84.
Tsai, Jung-hsiang. 2011. "Coalition Termination and Political Instability in Semipresidentialism," What Ever Happened to North-South?, IPSA-ECPR Joint
Conference, Brazil, Sao Paulo, February 16 to 19.
Warwick, Paul. 1992. “Ideological Diversity and Government Survival in Western
European Parliamentary Democracies,” Comparative Political Studies 25(3):
332-61.
15