ArticlePDF Available

Four Rounds in a Row: The Impact of Présidential Election Outcomes on Legislative Elections in France

Authors:

Abstract and Figures

This article focuses on the consequences of presidential election outcomes on legislative election results in the context of the French semi-presidential regime. Through an analysis based on aggregated national- and constituency-level data, it shows that presidential elections do have an impact on legislative elections. Furthermore, this impact is proven to affect the balance between presidential majority and opposition coalitions and between core and fringe parties. The 2000 and 2002 reforms aligning presidential and National Assembly mandates and instituting a situation of repeated honeymoon elections significantly decrease the likelihood of any future period of cohabitation between a president and an Assembly from opposing partisan coalitions. They also foster presidentialization and explain the relative decrease in fragmentation within the party system.
Content may be subject to copyright.
Original Article
Four rounds in a row: The impact of presidential
election outcomes on legislative elections in France
Elisabeth Dupoirier
a
and Nicolas Sauger
b,
*
a
Sciences Po CEVIPOF, 98 rue de l’Universite
´
, F-75007 Paris, France.
E-mail: elisabeth.dupoirier@sciences-po.fr
b
Sciences Po CEE, 98 rue de l’Universite
´
, F-75007 Paris, France.
E-mail: nicolas.sauger@sciences-po.fr
*Corresponding author.
Abstract This article focuses on the consequences of presidential election
outcomes on legislative election results in the context of the French semi-
presidential regime. Through an analysis based on aggregated national- and
constituency-level data, it shows that presidential elections do have an impact on
legislative elections. Furthermore, this impact is proven to affect the balance
between presidential majority and opposition coalitions and between core and
fringe parties. The 2000 and 2002 reforms aligning presidential and National
Assembly mandates and instituting a situation of repeated honeymoon elections
significantly decrease the likelihood of any future period of cohabitation between a
president and an Assembly from opposing partisan coalitions. They also foster
presidentialization and explain the relative decrease in fragmentation within the
party system.
French Politics (2010) 8, 21–41. doi:10.105 7/fp.2009.41
Keywords: election timing; France; electoral cycle; honeymoon/midterm elections;
semi-presidentialism
Introduction
On 24 September 2000, a referendum approved President Chirac’s proposal to
reduce the French presidential term from 7 to 5 years. Despite an apparent
agreement on this constitutional change, the actual support for it was less
clear-cut. The large majority (73.2 per cent) of people who voted ‘yes’
represented a mere 18.5 per cent of all registered voters. Turnout reached
the lowest level ever in the Fifth Republic, at only 30.19 per cent, and more
than 16 per cent of the ballots were spoiled. Indeed, the political elite itself
r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
www.palgrave-journals.com/fp/
was rather ambivalent about this reform. Even if a broad consensus was
achieved on the principle of the reform, the initiative itself came from outside:
Vale
´
ry Giscard d’Estaing, a former French president, put the issue on the
political agenda. Jacques Chirac’s prime minister in a period of cohabitation’,
the socialist Lionel Jospin, endorsed the idea, which came from one of the
president’s greatest enemies. Jacques Chirac triggered the referendum because
he could not see how to oppose such an idea and/or because he thought it could
help his plan to stand during the next presidential election by calling for a
second but shorte ned term.
The decision had tremendous consequences for the semi-presidential regime
of the Fifth Republic. It brought the mandates of both the National Assembly
and the president into alignment. Furthermore, in 2002, these terms coincided
for the first time. The legislative election was to be held just a few weeks before
the presidential election. However, in a strategic move, Lionel Jospin decided
to postpone the legislative election by 2 months, so that it followed the
presidential election by 4 weeks. He believed that his only chance of securing a
socialist majority in the National Assembly depended on his election as
president (Je
´
roˆ me et al, 2003).
This article focuses on the expected consequences of these two major shifts in
the French institutional model. Our main assumption is that both changes led
to the actual transformation of the working of the political system, because of
its presidentialization and the transformation of the pa rty system into a quasi
two-party system. As cohabitation has been a major issue in French politics
(Parodi, 1997; Elgie, 2002) since the 1980s, and because the 2000–2002 reforms
explicitly aimed at avoiding future cohabitation as far as possible, we examine
the likelihood that a ‘newly elected’ president is certain to be supported by a
congruent majority in the National Assembly.
The question is linked to the wider theory of the timing of elections and
electoral cycles
1
(Shugart and Carey, 1992; Smith, 2004) that provide the
general theoretical framework of this article. More precisely, we build on
Shugart and Carey’s (1992) analysis of the interactions among outcomes of
legislative and presidential elections in presidentia l or semi-presidential
regimes. The hypothesis is primarily that the relative timing of these elections
is important in assessing the likelihood of any form of divided government
emerging. If both elections are held in a short period of time, it is likely that
both the President and the majority in Parliament will be the same. There are
two main explanations for this: first, the two elections are held in the same
context and are hence more likely to produc e the same outcome; second, the
result of the first election in time (and possibly the anticipation of the outcome
of the second election) has consequences for the subsequent election because of
bandwagon dynamics, spillover effects and the fact that voters may have
preferences as to the grouped result of both elect ions rather than considering
Dupoirier and Sauger
22 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
each election independently (that is voters may, for instance, want to avoid a
situation of divided government).
2
To analyse the effects of the outcomes of presidential elections on the
legislative elections, this research is based on the aggregated results of French
elections collected at the national and constituency levels, since the first direct
presidential election in 1965. Over the last five decades of the French Fifth
Republic (before and after the 2000 and 2002 electoral calendar reforms),
almost all possible configurations of tim ing between presidential and legislative
elections have occurred. Despite a rather limited number of cases (only eight
presidential elections), we contend that France provides a unique set of data to
test for the effects of election timing and to discover steadfast rules prescribing
how presidential outcomes feed legislative dynamics. Our study of course takes
into account the fact that the party system itself has changed between the
sixties and the present (Sauger, 2009).
The remainder of this article is structured as follows. The second section
deals with the two institutional reforms that have de facto linked presidential
and legislative elections. The third section tracks shifts in electoral performance
for various parties with respect to the different cycles. The fourth section
looks at the effects of local logic on legislative elections. It asses ses the impact
of territorial dispersion on the parties’ performance. The fifth section outlines
a general model for forecasting legislative results for the parties from the
outcomes of the first round of the presidential election. The sixth section
concludes on the relative influence of the different effects created by the
2000–2002 reforms and proposes some hypotheses for the effects of election
timing.
National Elections and French Institutions
French electoral rules for presidential and legislative elections
As a semi-presidential regime, the French Fifth Republic is characterized by
the coexistence of direct presidentia l and legislative elections. With the
exception of the 1986 legislative election, presidential and legislative elections
are based on almost identical electoral formulas that belong to the large
family of majority systems. French presidents are elected through a majority-
runoff system that organizes a second and final ballot if no candidate obtains
a majority after the first ballot. The runoff ballot is held between the two
candidates who receive the highest number of votes on the first ballot.
Legislative elections follow a two-round majority plurality system with
a second and final ballot if no candidate obtains a majority on the first ballot.
A threshold of 12.5 per cent of registered voters decides who among the
Impact of presidential election outcomes in France
23r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
first round of candidates will be allowed to stand during the second
round.
3
The winner is the candidate who secures a plurality of the votes.
The main consequence of majority-runoff electoral systems, at least in the
French context, has been the bipolarization of the party system (Parodi,
1997).
4
From 1965 to 2002, the system experienced only three periods of
cohabitation
5
for a total of 9 years out of 37 in the entire period. In fact, this
can also be explained by the overall electoral dominance of the right from the
1960s to the late 1970s. Cohabitations quickly appeared after the first alter-
nation in office in 1981.
The 2002 reform led to a crucial re-ordering of the electoral cycle, so that
legislative elections now follow presi dential elections. Otherwise, ‘counter-
honeymoon’ elections, as Shugart and Carey (1992) label them, would have
fostered the parliamentary character of the regime rather than speeding up
its presidentialization. Yet, the ne w calendar introduced by the 2002 reform
has not been ‘constitutionalized’. This means that the two types of election may
be once again disconnected in the future if the president should die, for
instance. However, such a backward shift remains unlikel y in the long run, as
long as the president maintains the power to break up the Assembly, and as
long as the presidentialist reading of the constitution persists.
The impact of election timing and the significance of the 2000 and 2002 reforms
The question of election timing introduced by the 2000 and 2002 reforms
cannot be directly compared to the British situation. In his study of the
Westminster system, Alastair Smith shows that early British general elections
are uncommon, and might be considered as a signal leaders lack confidence in
their future performance (Smi th, 2004, p. 5). The first two French cases of early
legislative elections (the 1981 and 1988 dissolutions) were not the result of such
pessimism. They aimed to ensure that the presidential and legislative majorities
would be from the same political camp at the very beginning of the new
presidential term, so that the newly elected president would be provided with
a steady working majority to govern with.
The president’s power to call a legislative election for strategic reasons has,
however, seldom been used in France. In 1974, for instance, Vale
´
ry Giscard
d’Estaing, who had just been elected president, decided to accept and cope with
an Assembly dominated by the Gaullist parliamentary group rather than
dissolve the Assembly.
6
In 1995, Jacques Chirac also promised his followe rs
that he would not dissolve the Assembly. Chirac used this as an argument to
convince the RPR and UDF deputies to campaign for him once polls made
him a credible challenger against the incumbent Gaullist Prime Minister,
Dupoirier and Sauger
24 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Edouard Balladur.
7
Finally, the National Assembly has only been dissolved
four times: in 1968, because of the so-called e
´
ve
`
nements de mai; in 1981
and 1988 after the election of Franc¸ois Mitterrand as president because
the right was dominant in the previously elected National Assembly; and in
1997.
This 1997 case is in fact the only time the president has called for a legislative
election before the due date, using his power of dissolution to shape the
electoral agenda much as British prime ministers are used to doing. Indeed, the
1997 dissolution was viewed as a strategic decision by President Chirac
designed to provide his own partisan coalition with the best chance of winning
the forthcoming elect ion. It was also seen as the means by which he could gain
a firmer grasp on his own camp at a time when his leadership was clearly under
question. The decision was taken nearly 2 years after Chirac’s election as
president, at a moment when his presidential legitimacy was already tempered
(Carcassonne, 1995). However, the plan backfired and Chirac’s fatal error in
calling for a legislative election eventually led to the return of a left wing
majority in the new Assembly.
As analysed by Shugart and Carey (1992),the question of the political use of
election timing is much more relevant for the French case. They distinguish
honeymoon legislative elections, held during the first year of a new presidentia l
term, from midterm elections held afterwards. Based on a thorough
comparative analysis, the two authors conclude by underlining the divergent
political effects of honeymoon or midterm elections: honeymoon elections are
more likely than other electoral cycles to produce an Assembly majority for the
just-elected president, whereas midterm elections go the farthest in weakening
the impact of presidential elections over elections to the Assembly (Shugart
and Carey, 1992, pp. 263–265). These contrasting consequences are explained
by a number of overlapping logics, which are not specifically explored in this
article.
To come back to the French case, the 2000 and 2002 reforms obviously
foster the presidentialization of the regime (Poguntke and Webb, 2005), as the
president’s election inaugurates the whole process of allocating power. We
argue that newly elected presidents can rely on the election of a working
majority in the National Assembly 2 months after their own election, assuming
that the president successfully campaigns for the partisan coalition that has just
supported her/him. As Cox points out, within individual legislative constitu-
encies, voters and parties need to coalesce in order to convert their votes into
legislative seats more efficiently (Cox, 1997, p. 203). We argue that the newly
elected president is the right person to help the new partisan coalition win at
both the local and the national level: s/he is the most prominent political leader
to provide nationwide support to build the legislative coalition s/he needs to
support the incomi ng governm ent.
Impact of presidential election outcomes in France
25r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Election Timing and Its Political Effects on Legislative Elections
A classification of electoral cycles based on election timing
Given that France is a semi-presidential system with both direct legislative
and presidential elections, we propose here to map the configurations under
which the allocation of power has been processed. Eleven electoral cycles
8
have
been formed by presidential and legislative elections throughout the five
decades of the history of the Fifth Repu blic with respect to the period of time
that separates a legislative election from the previous presidential election (see
Table 1).
Table 1 shows the three types of electoral cycles based on the election timing
that we have delineated. The first type, with honeymoon legislative elections,
includes four cases that took place before and after the 2000–2002 reforms.
The second type of electoral cycle concerns midterm legislative elections
and includes the same number of cases as the first type. As a result of the use
of presidential power to dissolve the Assembly and the different lengths of
presidential and legislative terms before 2000, a third type of electoral cycle
should also be pointed out. There are only three cases of these and each cycle
began and ended with a legislative election. Within this third type, the first
cycle began with the just-in-time 1967 election and ended with the unexpected
1968 election held after General de Gaulle’s decision to dissolve the Assembly
with the hope of putting a stop to the ‘e
´
ve
`
nements de mai 1968’.
9
The other
two cycles of this type (1981–1986 and 1988–1993) began as a result of the
Table 1: Three types of electoral cycles from 1965 to 2007
Types of
electoral cycles
With honeymoon
legislative elections
With midterm
legislative elections
From one legislative
election to the next
Between 1965
and 2000
K 1981 P and L K 1965 (P) to 1967 (L) K 1967 (L) to 1968 (L)
K 1988 P and L K 1969 (P) to 1973 (L) K 1981 (L) to 1986 (L)
K 1974 (P) to 1978 (L) K 1988 (L) to 1993 (L)
K 1995 (P) to 1997 (L)
From 2000
to nowadays
K 2002 P and L
K 2007 P and L
Number of
cases
Four cases: Four cases Three cases
Two before the 2000 law
Two after the 2000 law
Abbreviations: P=Presidential election; L=Legislative election.
Note: The 1967–1968 cycle is excluded from any subsequent analysis because of its peculiarity.
Dupoirier and Sauger
26 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
dissolution of the Assembly following the election of President Mitterrand
(1981 and 1 988). They ended with a just-in-time midterm legislative election,
and led to periods of cohabitation. This third category of cycle is indeed
heterogeneous. The former can be seen as exceptional, whereas the two latter
cycles are in fact what is generally considered cycles ending with midterm
elections. Because the 1967–1968 cycle is exceptional, it is excluded from
any subsequent analysis. The 1981–1986 and 1988–1993 cycles are here
considered as a ‘control group’. The premise of this article is to look at the
transformation of electoral outco mes from presidential elections to legislative
elections; in this regard, these two cycles make it possible to more precisely
assess the specific impact of the presidentia l elections and the differences
between presidential and legislative elections beyond changes in the context
or the stake of elections.
The second type of electoral cycle, including midterm legislative elections, is
more interesting than it might at first appear to be. Three of the four cases
began or ended with an unexpected call for elections. Neither the 1965 nor the
1969 presidential election was planned when the early legislative elections took
place: that in 1965 was the result of a successful referendum that ushered in the
constitutional reform that provided for the direct election of the president; that
in 1969 was the result of President de Gaulle’s resignation after the failure
of the referendum on reforms to the French state.
10
Finally, the election in 1974
was the result of President Pompidou’s unexpected death.
Moving on from this classification of electoral cycles, we will now examine
the extent to which election timing affects legislatives outcomes.
The political impact of election timing on legislative elections
Let us now assess the extent to which the 2000 and 2002 reforms introduced a
seminal disruption in the way the French electoral system works and whether
these reforms have altered the structure of competition between parties.
In order to begin an an alysis of the relationship between presidential results
and honeymoon or midterm legislative elections, Figure 1 plots the national
results for each party or partisan family after the presidential election or the
first legislative election for the control group (x-axis) and after the subsequent
legislative election (y-axis).
11
In order to take the fragmented design of the
French system of parties into account, we have worked at the level of parties
such as the PCF and of families of parties. This notion of families of parties
describes aggregations of parties based on ideological or political proximities at
a precise political moment. It should be underlined that those parties are not
necessarily allied, as is the case for a party coalition, and that the scope of a
family may shif t from one election to another.
12
Impact of presidential election outcomes in France
27r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
The results shown by this graph are quite clear-cut. As expected, legislative
elections are not independent of the outcomes of the preceding presidential
election even in the case of midterm elections. For midterm legislative elections,
core parties have poorer performances while the regression line for honeymoon
elections is far steeper. For instance, a family of parties representing 40 per cent
of the votes in the presidential election is expected to win 43 per cent in the case
of a honeymoon election but only 34 per cent in a midterm election. In other
words, core parties benefit from honeymoon elections whether or not they are
part of the presidential majority, whereas smaller parties benefit from midterm
elections. This result in fact confirms previous analyses showing that the
fragmentation of the party system is greater in the case of midterm elections
and is minimal in honeymoon elections (Sauger, 2009). The ‘control group’
(which is constituted by groups of two legislative elections and close to the
notion of midterm elections and has thus to be compared to the ‘midterm
group) shows that presidential elections amplify the width of electoral changes.
Despite the fact that the ‘control group’ includes the historical defeat of
the Socialist Party (PS) in 1993, the regression coefficient between the results of
the two elections of this group is significantly higher when compared to the
‘midterm’ group.
0
10
20
30
40
50
60
Presidential election
Legislative election
Midterm
Honeymoon
Control
Honeymoon (linear)
Midterm (linear)
Control (linear)
01020
30
40 50 60
Figure 1: Electoral results for party families in the two first rounds of electoral cycles.
Notes: Regression results (OLS regression, all coefficients are significant at .01 level):
K Midterm: y ¼ 0.835x; R
2
¼ 0.698 (if three outliers are dropped, y ¼ 0.829; R
2
¼ 0.919)
K Honeymoon: y ¼ 1,06x; R
2
¼ 0.897
K Control: y ¼ 0.893x; R
2
¼ 0.825.
Dupoirier and Sauger
28 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Table 2 provides more detailed information on this relationship between two
elections belonging to the same electoral cycle. It presents the correlations
between the presid ential and legislative scores of parties (or families of parties)
at the constituency level. The values of Pearson’s r coefficients computed at the
local level
13
for each partisan family (or party) and for each couple of
presidential and legislative elections show the interdependent relationship
between the outcomes of the two elections, irrespective of the type of electoral
cycle they belong to. Three precise conclusions may be drawn from Table 2.
First, if we consider two elections belonging to the same electoral cycle, the
value of the Pearson coefficients varies in accordance with the nature of each
party or family of parties whatever the couple of national elections might be.
Fringe parties such as the Communist party (PCF) or the Extreme Right Wing
family (ExR) tend to benefit from steady geographical and political structures.
Therefore, they are not very sensitive to the issue of election timing.
Second, the Right Wing (RW) family, structured by the Union pour
la De
´
mocratie Franc¸aise (Union for French Democracy) (UDF) and
Rassemblement pour la Re
´
publique (Rally for the Republic) (RPR) in the
past and the Union pour la Majorite
´
Pre
´
sidentielle (Uni on for the Presidential
Majority) (UMP) at present, is much more sensitive to shifts in the partisan
system throughout the period. Until 1978 and Giscard d’Estaing’s decision
to found a party the UDF bringing together all the smaller move-
ments belonging to the non-Gaullist or centre-right families, the (RW) family
was weakened by rivalries between the centre and the RW. The values of the
Pearson correlations for the 1965/1967 electoral cycle reached 0.64 when
General de Gaulle himself represented the RW in the 1965 presidential election.
But the Pearson coefficient collapsed down to 0.28 during the next 1967/1969
electoral cycle when two candidates belonging to the same RW family both of
whom had an equally high level of notoriety
14
competed in the 1969
presidential election. However, since the 1978 legislative election, Pearson’s
r-values for the RW family of parties have always exceeded 0.80 whatever the
type of electoral cycle. This indicates the great ability of the RW family to
mobilize its composite electorate again whatever the political context of the
national elections might be. It should be underlined that for legislative elections
between 1978 and 2002, the UDF and RPR succeeded in proposing a sole and
common candidate for both parties in each constituency, with only a small
number of exceptions.
Third, unlike the RW family of parties, the ability of left wing parties or
party families to hold on to voters from one election to the next seems to be
much more dependent on the type of electoral cycle concerned. In the case of
midterm legislative elections, correlations between the PCF scores at presi-
dential and legislative elections reach or surpass 0.75, despite the fact that for
both the 1965 and 1974 presidential elections the PCF chose to endorse
Impact of presidential election outcomes in France
29r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Table 2: Correlations between scores obtained by party families from presidential to subsequent legislative elections at constituency level
Type of electoral cycle Electoral cycles ExL PCF NCL Centre RW ExR N cases Pres majority Leg majority
With honeymoon legislative elections P81/L81 0.08 0.94 0.77 0.95 474 L L
P88/L88 0.31 0.92 0.78 0.89 0.92 555 L L
P2002/L2002 0.49 0.76 0.57 0.86 0.92 555 R R
P2007/L2007 0.38 0.77 0.73 0.68 0.83 0.76 555 R R
With midterm legislative elections P65/L67 0.75
a
0.49
b
0.54 0.64 0.72 470 R R
P69/L73 0.52 0.94 0.52 0.29 0.28 473 R R
P74/L78 0.10 0.84
a
0.28
b
0.91 0.16 474 R R
P95/L97 0.35 0.89 0.70 0.93 0.92 555 R L
a
Correlation Mitterrand 1965 or 1974 / PC 1967 or 1978.
b
Correlation Mitterrand 1965 or 1974/ PS 1967 or 1978.
Abbreviations: Pres=Presidential election; Leg=Legislative election; L=Left wing majority; R=Right wing majority.
For the names of the columns see footnote 13.
Dupoirier and Sauger
30 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Franc¸ois Mitterrand’s candidacy rather than to present its own candidate.
Before Franc¸ois Mitterr and built a new party the PS the old socialist
movement the Section Franc¸aise de l’Internationale Ouvrie
`
re (French section
of the workers’ internati onal organization) (SFIO) which provided most of
the non-Communist candidates, was a parochial party that gave a great deal of
autonomy to its grassroots activists in the choice of legi slative candidates. The
political heterogeneity of those candidates for the 1967 legislative election goes
a long way in explaining the weak value of the Pearson coefficients. Comp are
the 1967 legislative election results to the 1965 presidential election results
(0.49) when Mitterrand declared himself as the candidate supported by all the
leftist movements. For a long time after the foundation of the PS, the
relationship between the presidential and legislative structures of the Non-
Communist Left family (NCL) remained unsteady and weak. They were always
weaker than those of the RW family: the value of the Pearson coefficient was
0.52 for the 1969–1973 cycle and collapsed to 0.28 for the 1974–1978 cycle, even
though the left wing allied parties
15
had begun to coordinate with the per-
spective of Mitterr and’s presidential victo ry in 1981. Therefore, M itterrand’s
decision to dissolve the Assembly after his presidential victories of 1981 and
1988 provided the opportunity to his party the PS to benefit from his own
success at the expense of its closer allies: Pearson coefficient values reached
0.77 and 0.78, and remain the highest values the PS has ever reached.
The differences in correlations between the RW family on the one hand and
the NCL family on the other are partly explained by the more or less volatile
nature of each block throughout the 1965–2007 period. During this time,
correlations were rather low for the RW family, while the centre (which was
actually more a centre-right group) had significant electoral results, as in the
1965/1967 and even the 1969/1973 electoral cycles. The NCL family, and
especially the PS, appear to be much more fragile than their (RW) UMP
challenger throughout the period.
There are several reasons for this situation. The main explanation is the
fragmentation of the NCL family. Given the fact that the PS alone is unable
to reach a majority of seats in the Ass embly, it has needed to build electoral
and governmental alliances for many years now. These have to be renewed
after each new electoral appointment, given that all the PS leftist partners are
weak (such as the PCF today) or weakly structured movements without solid
national organizations (such as the Greens). Or indeed they may be local
notables who seek support from the PS to increase their chances of being elected.
The second reason is that the PS has gradually lost its close relationship
with the working classes and has become one of the preferred parties of middle-
class voters, who are among the most volatile social categories of voters.
The PS therefore currently suffers from the tendency of those middle- (and
sometimes upper-) class voters to cast their vote in favour of attractive
Impact of presidential election outcomes in France
31r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
fringe parties. The shaky relationship between the PS and its partners might
be considered a structural weakness as compared to the UMP’s capacity to
organize its own camp.
These results confirm the significance of the timing of elections in forecasting
the results of legislative elections. Midterm and honeymoon elections are both
closely related to the outcomes of the preceding presidential election but to the
benefit of different parties or party families depending on whether the election
is a honeymoon or a midterm one. However, the effect of election timing is not
sufficient to understand the shift in the structures of political competition from
the presidential to the legislative elections. A closer look at the logic of the
legislative elections is therefore necessary to assess the embedded impacts of the
timing and territorial distribut ion of electoral competition.
From Presidential Success to a Legislative Working Majority
In contrast to presidential elections based on a single configuration of
candidates for a single national constituency, legislative elections are held
with a great diversity of cand idate configurations. In theory, these can be as
numerous as the number of constituencies in which deputie s are elected. Not all
parties that have just supported a presi dential candidate put forward a
candidate in each constituency as a result of financial or organizational
difficulties or sometimes local or national trade-offs. For instance, since 1997,
the PCF has no longer been in a position to put a candidate forward in each
constituency.
16
In contrast, some parties that are not in a position to put a
presidential candidate forward or to have legislative candidates everywhere
nevertheless present candidates in some constituencies as a result of resources
provided by local grassroots activists.
17
That is to say, for a newly elected
president the probability of reaching a working majority of seats in the new
Assembly in order to initiate new legislation depends on the final balance
between 577 local outcomes.
18
With respect to this assumption, each
constituency has a specific political identity that prevents the outcomes of
the legislative election from exactly replicating the national political tendency
confirmed by the presidential election. We therefore take this situation into
account as a ‘local variable’, and argue that it is powerful enough to affect the
outcomes of each constituency to a greater or lesser degree, all the more so as
the power of this variable differs between honeymoon and midterm elections.
19
The question is now that of whether the influence of the local variable is
powerful enough to significantly modify the score of a party from one
constituency to another. The results shown in Table 3 confirm, first, that the
influence of the local variable is much more significant for legislative than for
presidential elections.
Dupoirier and Sauger
32 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Second, core partisan families are much more sensitive to the effects of the
local variable than fringe partisan families. For legislative elections, their
territorial variations may reach more than 10 times the value of the standard
deviation in the case of the RW family and almost 10 times the standard
deviation in the case of the PCF or the NCL family. For electoral cycles with
honeymoon legislative elections, increases in standard deviations from the
presidential election to the subsequent legislative election are impressive:
approximately þ 4 points for each partisan family.
As has just been suggest ed, the influence of the local variable varies from one
party to another. According to our research question, the most important
result is that the current institutionalization of electoral cycles with honeymoon
legislative elections reinforces the territorial dispersion of the results of
legislative elections. It therefo re balances the national political influence of the
newly elected presidential coalition on the legislative election. Compared to
electoral cycles with midterm elections, the value of the standard deviation for
each core party or party family increases significantly: from 2.4 to 4.4 for the
PCF, from 3.0 to 4.0 for the NCL coalition, and from 0 to 3.7 for the RW
coalition.
Obviously, the local variable provides some autonomy to legislative elections
with respect to presid ential influence. Unlike presidential elections, which are
ruled by a national institutional issue, thus weakening territorial political
disparities, legislative elections are much be tter balanced between national and
local issues. When legislative elections do not carry a national political message
that has just been addressed by means of the presidential election (as is the case
for honeymoon elections), the influence of local variables on the outcomes of
the election would reach a maximum were it not hidden behind a low turnout.
In contrast, when legislative elections have a national message to deliver to the
president and his government, as midterm elections do, local disparities are
Table 3: Standard deviation of scores within party families with respect to election timing
ExL PCF NCL RW ExR
Electoral cycle with honeymoon legislative elections
Average standard deviation for presidential elections 2.4 3.6 4.9 6.8 4.7
Average standard deviation for legislatives elections 1.3 8.0 8.9 10.5 4.1
Difference between presidential and legislative elections 1.1 þ 4.4 þ 4.0 þ 3.7 0.6
Electoral cycle with midterm legislative elections
Average standard deviation for presidential elections 2.9 6.2 6.1 10.7 2.0
Average standard deviation for legislatives elections 2.3 8.6 9.1 10.7 2.8
Difference between presidential and legislative elections 0.6 þ 2.4 þ 3.0 þ 0.8
Impact of presidential election outcomes in France
33r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
overtaken by the national issue central to the election, added to the fact that
turnout rises.
Modelling Legislative Election Performance from Presidential Outcomes
It is clear from the previous analyses that the outcome of legislative elections
is linked to the outcome of the preceding presidential election but also
depends on the timing of the electoral cycle. This section systematizes these
observations by modelling the results of legislative elections from a limited set
of explanatory variables chosen to describe both the outcomes of the preceding
elections and the type of electoral cycles to which the legislative elections
belong.
The analysis strategy is rather straightforward. Using a simple OLS model,
we regress the result of each party family or party in the first round of
a legislative election
20
on the results of the preceding presidential election;
the difference in turnout between presidential and legislative elections; the
winner of the presidential race; and the type of legislative election involved:
honeymoon or midterm.
21
In addition, we hypothesi ze that the effect of
midterm or honeymoon elections on the core parties is variable according
to their affiliation or opposition to the presidential majority coalition. Using
two dummies, the regressions also take into account the interaction between
the type of legislative election and support for the elected president: midterm
majority and honeymoon majority. We also added a control variable to
allow for the effect of turnout variation from the presidential election to the
subsequent legislative election. Another control variable has been added
through the score obtained by the parties during the preceding legislative
election. Despite obvious co-linearity problems with the results of the presi-
dential election, this variable nonetheless shows interesting results, consistent
with models in which it is dropped. The regression model takes thus into
account all dimensions discussed earlier in this paper. The descriptive statistics
of all metric variables in this model are presented in Table 4.
We ran these regressions for the period 1973–2007, with the exception of the
1978 elections. The first and third electoral cycles are excluded because the
PCF did not run any candidate for the 1965 or the 1974 presidential elections.
We ran specific regressions for the PCF and the ExR family for a more
restricted period (1988–2007), so that the level of support for these parties
would be coherent throughout the period. Table 5 shows the results of the five
regressions.
The first striking result is that the outcomes of the legislative elections
subsequent to a presidential election depend more on the results of the previous
legislative election than the results of the prior presidential election whatever
Dupoirier and Sauger
34 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Table 4: Descriptive statistics of electoral results by party grouping
RW NCL ExL PCF (1988–2007) ExR (1988–2007)
Min Max Mean Min Max Mean Min Max Mean Min Max Mean Min Max Mean
Legislative score 0.20 0.85 0.4191 0.01 0.71 0.3134 0.00 0.64 0.1289 0.00 0.49 0.0842 0.00 0.38 0.1126
Presidential score 0.16 0.92 0.4215 0.02 0.48 0.2681 0.02 0.54 0.1448 0.00 0.29 0.0646 0.03 0.34 0.1498
Previous legislative election score 0.20 0.88 0.4318 0.01 0.69 0.2973 0.00 0.60 0.1411 0.00 0.47 0.0933 0.00 0.38 0.1229
Difference in turnout 0.45 0.28 0.1193 0.45 0.28 0.1193 0.45 0.28 0.1193 0.45 0.08 0.1463 0.45 0.08 0.1463
N 3721 3722 3722 2775 2775
Impact of presidential election outcomes in France
35r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
Table 5: OLS regression modelling party scores for legislative elections (first round) depending on election timing (1973–2007)
RW NCL ExL PCF (1988–2007) ExR (1988–2007)
b sign b sign b sign b sign b sign
Constant 0.02 *** 0.00 0.08 *** 0.04 *** 0.03 ***
Presidential score 0.32 *** 0.52 *** 0.55 *** 0.5 *** 0.71 ***
Previous legislative election score 0.63 *** 0.56 *** 0.59 *** 0.69 *** 0.2 ***
Difference in turnout 0.09 *** 0.08 *** 0.07 *** 0.06 *** 0.04 ***
Midterm, opposition 0.04 *** 0.06 *** 0.04 *** 0.04 *** 0.05 ***
Midterm, majority 0.05 *** 0.09 *** 0.02 *** 0.01 ***
Honeymoon, majority 0.09 *** 0.06 *** 0.01 *** 0.02 ***
Honeymoon, opposition 0.04 *** 0.03 *** 0.04 *** 0.04 0.05 ***
R
2
0.74 0.71 0.87 0.85 0.86
Coefficients have to be multiplied by 100 to be read in percentage points.
*** P value o0.01
Note: The four variables Midterm/Honeymoon Opposition/Majority reflect the four possible positions for any party entering the electoral race.
Coefficient values are here capturing the ‘net’ effect of the position.
Dupoirier and Sauger
36 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
the party or the party family might be, with the singl e exception of the ExR
family. The case of the RW party family is the most surprising. The coefficient
of the previous legislative score is the highest, reaching 0.63, whereas the
presidential coefficient reaches only 0.32. If the score of the previous legislative
election is drop ped from this model, the coefficie nt for the presidential election
obviously increases. However, dropping this variab le leads the R
2
of the
regression to drop from 0.74 to a mere 0.58. This observation unde rlines the
significant role of the local variable. The political orientation of the incumbent
deputy, the level of the party’s competitiveness, and the strategies of the parties
or the candidates in each constituency as summarized in the local variable are
all significant components of the constituency’s history and contribute to the
shaping of results in the legislative election.
However, the logic of the previous election remains important: both
presidential and legislative coefficients are strongly positive and significant.
For all left wing pa rties, both previous elections are similar, especially in the
case of the NCL (0.52 and 0.56) and the Extreme Left (ExL) family (0.55 and
0.59). In contrast, legislative scores obtained by ExR candidates depend
essentially on Jean Marie Le Pen’s presidential scores (0.71), showing that the
Front national is in fact the only party capable of structuring this party family.
Since 1988, it has had a national strategy of fielding candidates in every
constituency, thus presenting a coherent supply throughout the territory. In
addition, the personality of Jean Marie Le Pen provides a key to understand ing
the success of the ExR at national level despite the fact that the FN is not
strongly rooted at the local level.
The significant role of the election timing variable is confirmed. Midterm
elections are always profitable for core parties outside the presidential majority.
They all gain consistently from 3.5 per cent to 5.6 per cent when they are
outsiders. In contrast, honeymoon elections favour parties belonging to the
presidential majority. This is especially the case for the RW party family, which
increases its score by 9 per cent in honeymoon elections, whereas the PCF
and the NCL party family experience much more limited gains (2 per cent and
6 per cent, respectively). Fourth, midterm elections are uncomfortable situations
for all parties that support the presidential majority, and the NCL party family
loses much more approximately 9 per cent of its previous score than the RW
party family, which loses only 5 per cent. Fringe parties seem less hampered, the
ExL family winning on average 2 per cent and the PCF 0.5 per cent.
In other words, the honeymoon-majority and the midterm-majority situations
are symmetrical, both having an effect on the balance between majority and
opposition blocks, as well as on the balance between fringe and core parties.
These results allow the important obstacle represented by midterm elections
for any majority to be underlined. To avoid any possibi lity of cohabitation,
the left wing coalition must be more than 10 per cent of votes ahead of the
Impact of presidential election outcomes in France
37r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
RW opposition after the presidential election (this prediction of course does
not take into account the effects of the electoral system and the geographic al
distribution of votes; as the electoral system is highly disprop ortional, these
estimates can be read thus as a maximum of the minimum required to secure
a majority). For the RW coalition, the necessary lead is limited to 5 per cent of
the votes. With such results, it is easier to understand why the RW block
remained in office for such a long time until the 1981 presidential election.
In the same way, it could be suggested that at present when neither a Left Wing
nor a RW party coalition is sure of dominating their opponents by 10 points at
least after a presidential election, the institutionalization of the honeymoon
legislative election seems to be the best way to stay in office for a full presi-
dential term.
Conclusion
At the end of this analysis, we argue that both the 2000 and 2002 reforms led to
a significant re-ordering of the electoral system and in this way to a crucial
reshaping of the way in which the French political system works. The literature
on election timing and the notion of electoral cycles were especially useful
to understand how and to what extent controlling the electoral calendar is
an effective means of controlling the balance of powers between the president
and the Assembly in a semi-presidential regime. In the French case, the
institutionalization of electoral cycles that begin with a presidential election
and are immediately followed by a honeymoon legislative election shifts the
balance in favour of the president for two main reasons. The first is that the
president’s own election inaugurates the whole process of sharing out powers
between the different bodies; the second is that with an early legislative
election, presidents do not need to prove their capacity to govern. Both reasons
increase the likelihood that a newly elected president will gain a working
majority in the Assembly.
The empirical tests realized throughout an extended period of time
(1965–2007) provide e xamples of various configurations of time between
presidential and legislative elections, and allow three conclusions to be drawn.
First, all honeymoon elections favour the party coalition that has just
supported the new president. The president is the right person to help a new
national party coalition to win at the local level of the 577 constituencies that
each elects its own MP.
Second, and more generally speaking, honeymoon legislative elections work
in favour of core parties rather than fringe mo vements, which are often not
solid enough to campaign for two elections occurring in such a short period of
time. Therefore, honeymoon elections limit the tendency of the partisan system
Dupoirier and Sauger
38 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
towards fragmentation afte r the first round of both the presidential and
legislative election.
Third, and regarding the question of whether or not the president is likely to
obtain a working majority in the new Assembly , the most important feature of
honeymoon elections is that they reinforce the territorial dispersion of party
scores. Therefore, the national political influence of the newly elected
presidential coalition could be counterbalanced by parties belonging to the
presidential opposition if their candidates benefit from a solid local imple-
mentation. This was the case for the 2007 legislative election when the PS,
whose candidate lost the presidential election, won more seats than expected
during the subseq uent legislative election.
To a large extent, it could be argued that the institutionalization of
honeymoon legislative elections contributes to the steadiness of the political
system. At a time when neither a left wing nor a RW presidential coalition is in
a position to face a midterm election with a comfortable likelihood of giving
the president a working majority of seats in the new Assembly, a honeymoon
election may be an effective means of stabilizing presidential power until the
end of the term of office.
Should newly elected presidents take it for granted that they will be
supported by a working majority in the National Assembly? The likelihoo d of
any cohabitation in the future has been reduced by the 2000 and 2002 reforms.
Nevertheless, as long as the presidential power to dissolve the assembly is
maintained, the possibility of a return of midterm legislative elections cannot
be excluded, given the fact that they are far less hazardous than a referendum
in solving a conflict between a president and some kind of extra parliamentary
opposition, for instance. The model may well be the 1968 midterm election
when de Gaulle used his power of dissolution to put an end to the e
´
ve
`
nements
de mai.
Notes
1 We describe an electoral cycle as a period of time opened by a national election presidential or
legislative and closed by the subsequent legislative election; this is a working definition for our
purpose taking into account the characteristics of the Fifth Republic.
2 See, for instance, Dolez and Laurent (2009), who propose a more thorough classification of
what people may have in mind in this kind of context.
3 This threshold has been in place since 1978. Therefore, the effect of the threshold is all the more
drastic when turnout is low. From 1958 to 1962, it was fixed at 5 per cent of the votes; from
1967 to 1973, the threshold was fixed at 10 per cent of registered voters.
4 The 1974 presidential election is viewed as the moment when the bipolarizations definitively
shaped the whole political system.
5 1986–1988; 1993–1995; 1997–2002.
6 Giscard was a member of a fringe Right Wing (RW) party allied to the powerful Gaullist party.
Impact of presidential election outcomes in France
39r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
7 Edouard Balladur, who was also a member of the RPR, was Franc¸ois Mitterrand’s prime
minister when the 1995 presidential election was held. Unlike Jacques Chirac, he did not say
anything during the campaign about the presidential power to break up the Assembly.
8 See footnote 1.
9 The Assembly elected in 1968 was a huge success for the RW parties that supported General de
Gaulle’s government.
10 Ge
´
ne
´
ral de Gaulle proposed the introduction of a new level of administration: the regional level.
11 All electoral results used in this article were collected from the French Ministry of the Interior’s
official results.
12 The NCL always includes the PS but other members come and go, such as the PSU, whose
candidates belonged to the Extreme Left Wing family until 1981 and then became members of
the NCL. Another example is the Verts (Greens) who belonged to the non-classified family until
1995 and then became members of the NCL family. Until 2002, the (RW) family included
different movements belonging to the Gaullist heritage, such as the RPR and movements
belonging to the more traditional members of French RW families such as the UDF. Since the
2002 legislative election, the RW family has been structured by the UMP party surrounded by
other weak movements. The Extreme Right Wing family (ExR) sometimes includes only Le
Pen’s party (the FN), but from time to time hosts the MNR party and/or other movements
ideologically similar to the FN.
13 By local level we mean the level of legislative constituencies. The Pearson coefficient assesses the
significance of the relationship between two series of data. Pearson’s R coefficients focus on the
structural relationship between two series of metric data without reference to how much each
piece of data scores. Therefore, a political force might be analysed over a long period of time,
irrespective of its tendency to increase (the Front national) or decrease (the PCF) throughout the
period.
14 Alain Poher was the President of the Senate and was acting president of the Republic after
President de Gaulle resigned; Georges Pompidou was the incumbent prime minister.
15 The allied parties were the PS, which was at the head of the coalition; the PCF as the more
powerful partner; and the radical party Mouvement des radicaux de gauche (Left radicals’
movement) (MRG), which was only significant in the South West of France. In addition to
some clubs, such as that founded by Mitterrand –the Convention des Institutions Republicaines
(CIR) those parties were linked by a political pact that was much more than an electoral
alliance, that is, a published governmental programme.
16 The number of communist candidates at legislative elections decreased from 97 per cent in 1997
to 92 per cent in 2007; before F. Mitterrand founded the PS, the socialist family was not able to
put forward a candidate in each constituency (only 87 per cent in 1967). In contrast, the FN,
which covered only 23 per cent of all constituencies in the 1981 legislative election, has been able
to put forward a candidate in each metropolitan constituency since 1988; one or several
candidates belonging to the Ex family has stood in each constituency since 2002. In contrast, the
NW partisan family put candidates forward throughout the entire period.
17 This is the case for ecologist movements throughout the period and among these parties, more
especially the Greens until they became members of the PS governmental coalition (1997).
18 The 577 seats in the Assembly are shared in the following way: 555 for metropolitan
constituencies and 22 for overseas constituencies.
19 This study is only based on the results of metropolitan constituencies, whose number has grown
from 474 to 555 (or 577 with overseas territories) since the mid-1960s.
20 This result is computed as a ratio of the number of valid votes cast.
21 The two variables (whether or not parties are part of the presidential majority, whether or not
elections are midterm) are in fact captured through three dummies in the regression (midterm,
midterm majority, honeymoon majority) that account for all possible cases.
Dupoirier and Sauger
40 r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
References
Carcassonne, G. (1995) La primaute
´
de l’e
´
lection pre
´
sidentielles. In: N. Wahl and J.-L. Quermonne
(eds.) La France pre
´
sidentielle: L’influence du suffrage universel sur la vie politique. Paris: Presses
de Sciences Po, pp. 31–45.
Cox, G. (1997) Making Votes Count: Strategic Coordination in the World’s Electoral Systems.
Cambridge: Cambridge University Press.
Dolez, B. and Laurent, A. (2009) Strategic voting in legislative majority runoff elections in
a semi-presidential system. The 2007 French legislative elections (mimeo).
Elgie, R. (2002) La cohabitation de longue dure
´
e: Studying the 1997–2002 experience. Modern &
Contemporary France 10(3): 297–313.
Je
´
roˆ me, B., Je
´
roˆ me-Speziari, V. and Lewis-Beck, M. (2003) Reordering the French election
calendar: Forecasting the consequences for 2002. European Journal of Political Research 42:
425–440.
Parodi, J.-L. (1997) Proportionnalite
´
pe
´
riodique, cohabitation et atomisation partisane: Un triple
de
´
fi pour le re
´
gime semi-pre
´
sidentiel de la Ve
`
me Re
´
publique. Revue Franc¸aise de Science
Politique 47(3–4): 292–312.
Poguntke, T. and Webb, P. (eds.) (2005) The Prezidentialization of Politics. Oxford: Oxford
University Press.
Sauger, N. (2009) The French party system: Fifty years of change. In: S. Brouard, A. Appleton, and
A. Mazur (eds.) The Fifth Republic at Fifty. London: Palgrave, pp. 79–98.
Shugart, M. and Carey, J. (1992) Presidents and Assemblies Constitutional Design and Electoral
Dynamics. Cambridge: Cambridge University Press.
Smith, A. (2004) Election Timing. Cambridge: Cambridge University Press.
Impact of presidential election outcomes in France
41r 2010 Macmillan Publishers Ltd. 1476-3419 French Politics Vol. 8, 1, 21–41
... The main feature of the French political system is its centralisation and the strength of the executive (Dupoirier and Sauger, 2010). The French parliament is thus considered a weak legislature (Costa and Kerrouche, 2009a). ...
... We focus here on written questions about agriculture submitted between 1997 and 2007. This period covers a time of significant changes, including the 2002 constitutional reforms that transformed the French political landscape by strengthening the link between the executive and the legislative bodies (Dupoirier and Sauger, 2010). This is also the period the literature identifies as the beginning of the trend towards the personalisation of politics (Thomassen and Andeweg, 2004). ...
Full-text available
Article
There is growing interest in personal representation and policy responsiveness in the European literature. The current scholarship mainly focuses on strong legislatures. This article aims to contribute to the discussion about policy responsiveness by investigating the least likely case, France. It asks whether French MPs engage in policy responsiveness, and if they do, how responsiveness interacts with party agenda and electoral vulnerability. Drawing on written parliamentary questions asked by French MPs from 1997 to 2007 ( N = 1172), the article shows that MPs’ behaviour is strongly affected by their district features. Moreover, the article shows that party agenda conditions the impact of district features. MPs are more likely to engage in policy responsiveness when the issue is also essential for the party. In other words, policy responsiveness is not independent of the party agenda. The article thus contributes to a better understanding of the personalisation of political representation and its relationship with political parties.
... The two rounds of presidential elections and the two rounds of legislative elections had coalesced into a single, four-round electoral sequence (Dupoirier and Sauger 2010;Parodi 2007). Following the presidential election, the "momentum" was such that the president's party was virtually guaranteed to win the legislative elections. ...
... It features a directly elected president who is both the head of state and government. And it goes hand in hand with a considerable personalization of politics, limiting political parties' influence (Dupoirier and Sauger, 2010;Lewis-Beck et al., 2012). As already mentioned, the synchronization of presidential and legislative electoral cycles has strengthened the centrality of the president (Grossman and Sauger, 2009). ...
... It features a directly elected president who is both the head of state and government. And it goes hand in hand with a considerable personalization of politics, limiting political parties' influence (Dupoirier and Sauger, 2010;Lewis-Beck et al., 2012). As already mentioned, the synchronization of presidential and legislative electoral cycles has strengthened the centrality of the president (Grossman and Sauger, 2009). ...
Chapter
This chapter presents a general theoretical framework to analyse party competition and its effects on policymaking. There are two perspectives: that parties still matter for policymaking despite signs of decline and that parties are driven by instrumental considerations and will do what it takes to get elected. This chapter adopts an intermediary perspective, arguing that campaign priorities not only reflect partisan preferences, but also respond to rival parties’ campaigns. This leads to an important common ground across different parties regarding their issue emphasis, i.e. the ‘tunnel’ of attention. Moving away from the heart of the tunnel is possible, but potentially costly. Post election, the victorious party will deliver on its programme priorities not least because of the pressure of junior coalition partners and the opposition. The tunnel of attention thus constrains party competition, but triggers mandate implementation.
... This was partly due to the reaffirmed right-wing cooperation through the UMP from 2002 onwards, but it owed far more to the realignment of the electoral calendar as part of the five-year presidential term (quinquennat) reforms two years previously. Since 2010, the primacy of the presidential race has turned the Assembly elections into essentially the third and fourth rounds of the presidential election (Dupoirier and Sauger 2010), with a steep fall-off in support for presidential losers and smaller parties. ...
Chapter
The chapter explains how the centrist Christian democratic Austrian People’s Party (ÖVP) has responded to the challenges of the Silent revolution and counter revolution by demonstrating a selective willingness to cooperate with the populist radical right Freedom Party (FPÖ). Polling and manifesto data show how the ÖVP has shifted from dismissive to accommodative approaches when it was polling behind the FPÖ. Sufficient electoral distress led to the installation of new leaders who were able to change the strategic status quo. In the first instance in 1995 Wolfgang Schüssel emphasized policy-seeking and in the second case Sebastian Kurz pursed vote-seeking. Both strategies resulted in a positional alignment and eventually a coalition with the FPÖ, which at the time was pursuing office. Changes in the ÖVP depended on shifts in the balance of power among important intra-party groups, specifically, hardline conservatives and market liberals viewing cooperation with the FPÖ as advantageous for their respective interests. Analysis of the supply side reveals the close programmatic alignment of ÖVP with FPÖ positions since 2002. Demand side analysis suggests that the programmatic shifts by the ÖVP coincide with changes in the profile of its electorate toward a composition more typical of a far right party. Overall, the chapter concludes that while the ÖVP has been affected by massive voter dealignment since the 1980s, it responded to the counter revolution and the resulting surge of nativist populism mainly by means of emulation and cooperation.
... Previous studies of the FN vote have noted a significant obstacle to party support from high abstention rates. Despite paradoxically strong partisan loyalty across electoral periods, evident in the 1990s (Shields, 1997) early 2000s (Mayer, 2002), and, after the 2007 switch to Sarkozy (Mayer, 2007) still evident at the commune level between the 2017 presidential first round and the 2019 European elections - Figure 2 panel A -second round performances have suffered from fall-off in turnout, and legislative elections, increasingly seen by voters as mere 'confirmatory' elections of the presidentials (Dupoirier and Sauger, 2010), have particularly seen this phenomenon in the wake of the RN presidential candidates' elimination. In the 2014 municipals, FN sympathisers were the second-most likely group of party supporters to abstain (Muxel, 2014: Table 2). ...
Full-text available
Preprint
The 2020 municipal elections represent a further stage in the institutionalisation of the Rassemblement National at the level of local governance, and at the same time underline the limits to the party's organisational capacity. Within a context of significant political and socioeconomic upheaval, the party's capacity to consolidate its 2014 victories, and build upon these through local mobilisation of political networks and exploitation of traditional opposition weaknesses, was limited to its existing strongholds in the NorthEast and South of the country. In particular, these successes were helped through the distancing of the local notables from national party labelling, and indicated the strength of pragmatic municipal governance in convincing voters to endorse candidates distancing themselves from the RN's extremist profile. Yet, only where a complex interplay of local socioeconomic factors, candidate profile and opposition weakness obtained could the party secure a limited number of victories.
Article
For the second time in the history of the Fifth Republic, the candidates of France’s former mainstream left- and right-wing parties were disqualified during the first round of the presidential elections. The repetition of the duel between Emmanuel Macron (centrist candidate) and Marine Le Pen (radical-right candidate) in the second round of the presidential election marks the durable transformation of the political space in French politics, an evolution that was already taking shape in 2017. The 2022 parliamentary elections confirmed the transformation and reconfiguration of the political landscape. Overall, the results of the elections in 2022 underscore major shifts in the party system as well as deep political fragmentations not only on the left but also, and in particular, on the right. Above all, the poor level of turnout in both elections reconfirms the considerable and enduring crisis of representative democracy in France.
Chapter
In spite of the fact that Conservative, Christian democratic and Liberal parties continue to play a crucial role in the democratic politics and governance of every Western European country, they are rarely paid the attention they deserve. This cutting-edge comparative collection, combining qualitative case studies with large-N quantitative analysis, reveals a mainstream right squeezed by the need to adapt to both 'the silent revolution' that has seen the spread of postmaterialist, liberal and cosmopolitan values and the backlash against those values – the 'silent counter-revolution' that has brought with it the rise of a myriad far right parties offering populist and nativist answers to many of the continent's thorniest political problems. What explains why some mainstream right parties seem to be coping with that challenge better than others? And does the temptation to ride the populist wave rather than resist it ultimately pose a danger to liberal democracy?
Chapter
Four institutional features such as confidence vote, electoral system, candidate nomination, and decree power can affect the degree of party discipline in France. Other than four institutional factors, this chapter also asserts that three contextual factors such as the popular support for the president, the ideological cohesiveness of the majority, and the majority micromanagement can affect the president’s ability to steer the legislative agenda. In France, the majority of presidents have experienced defections of partisan members who have different preferences or stances in conflict with theirs.
Full-text available
Article
Shows that the politics of democratic societies is moving towards a presidentialized working mode, even in the absence of formal institutional changes. These developments can be explained by a combination of long-term structural changes in modern politics and societies’ contingent factors that fluctuate over time. While these contingent, short-term factors relate to the personalities of office holders, the overall political agenda, and the majority situation in parliament, there are several structural factors that are relatively uniform across modern nations. First, the internationalization of modern politics (which is particularly pronounced within the European Union) has led to an ‘executive bias’ of the political process that has strengthened the role of political top elites vis-à-vis their parliamentary groups and/or their parties. Their predominance has been amplified further by the vastly expanded steering capacities of state machineries, which have severely reduced the scope of effective parliamentary control. At the same time, the declining stability of political alignments has increased the proportion of citizens whose voting decisions are not constrained by long-standing party loyalties. In conjunction with the mediatization of politics, this has increased the capacity of political leaders to bypass their party machines and to appeal directly to voters. As a result, three interrelated processes have led to a political process increasingly moulded by the inherent logic of presidentialism: increasing leadership power and autonomy within the political executive; increasing leadership power and autonomy within political parties; and increasingly leadership-centered electoral processes. The book presents evidence for this process of presidentialization for 14 modern democracies (including the USA and Canada). While there are substantial cross-national differences, the overall thesis holds: modern democracies are increasingly following a presidential logic of governance through which leadership is becoming more central and more powerful, but also increasingly dependent on successful immediate appeal to the mass public. Implications for democratic theory are considered.
Full-text available
Article
French politicians sometimes change election rules for political advantage. In the Spring of 2001, the ruling Socialists pushed through the inversion of the 2002 election calendar despite stiff opposition. What were the consequences of scheduling the presidential election before the legislature elections? Employing new techniques for French election forecasting, we show that the inversion brought great vote gains to the Socialist-led coalition in both the legislative and presidential arenas. One advantage of this forecasting methodology is that it allows counterfactual comparisons in advance of the actual election contest. Comparing the scenarios ‘inversion’ versus ‘no inversion’, the Socialist leadership appeared highly strategic and successful.
Book
This chapter focuses on the evolution of the French party system.
Book
Popular elections are at the heart of representative democracy. Thus, understanding the laws and practices that govern such elections is essential to understanding modern democracy. In this book, Cox views electoral laws as posing a variety of coordination problems that political forces must solve. Coordination problems - and with them the necessity of negotiating withdrawals, strategic voting, and other species of strategic coordination - arise in all electoral systems. This book employs a unified game-theoretic model to study strategic coordination worldwide and that relies primarily on constituency-level rather than national aggregate data in testing theoretical propositions about the effects of electoral laws. This book also considers not just what happens when political forces succeed in solving the coordination problems inherent in the electoral system they face but also what happens when they fail.
Chapter
Historically, both Belgium and the Netherlands are archetypes of ‘consociational democracies’. These are characterized by broad multi-party coalitions, numerous power-sharing devices, and fragile checks and balances in order to ensure due influence for all relevant parties and minority groups. Hence, the overarching logic of these consensus democracies seems to represent an obstacle to a process of presidentialization. However, we argue that the need for strong leadership resulted in more prominent and powerful positions for the (parliamentary) party leaders and Prime Ministers. We present evidence of a process of presidentialization that gained momentum a decade earlier in the Netherlands (from the 1970s onwards) than it did in Belgium (from the 1980s). It is interesting to note that the increased autonomy of Prime Ministers is not due to constitutional amendments, but tends to be linked to the increased decision-making role for the inner cabinet, the professionalization of the Prime Minister’s Office, and the growing attention the audiovisual media give to the Prime Minister. Similarly, parliamentary party leaders in The Netherlands and extra-parliamentary party leaders in Belgium grew stronger through an accumulation of power and resources at the leader’s office, personalized campaigning and a centralization of control over inner party selection procedures, and party leadership selection.
Article
Structure par le jeu combine des contraintes binaires de l'election presidentielle, du refe­rendum et du scrutin majoritaire a deux, tours et homogeneise au niveau electoral et parlementaire par la dissolution, le systeme semi-presidentiel francais a evolue au cours de ses deux premieres decennies vers un quasi-type-ideal qui culmine a la fin des annees soixante-dix. Mais il va devoir relever au cours des deux decennies suivantes trois defis majeurs -. Celui de la proportionnalisation periodique, qui fait alterner contrainte majoritaire et ouverture proportionnaliste, celui des cohabitations qui deplace le lieu du pouvoir et entraine la soumission tribunicienne du president, celui de l'atomisation partisane qui multiplie le nombre d'acteurs au premier tour et complique le jeu du second quand cer­tains d'entre eux n'entrent pas dans la logique des alliances. La dominante majoritaire du systeme mis en place en 1958-1962 semble cependant resister a ce triple defi. PERIODIC PROPORTIONALIZATION COHABITATION PARTY TRIPLE CHALLENGE FOR THE FRENCH FIFTH SEMI- PRESIDENTIAL REGIME
Article
Part I. Presidential and Parliamentary Democracy: 1. Basic choices in democratic regime types 2. What is presidentialism? Criticisms and responses 3. The constitutional origin and survival of assembly and executive 4. Legislative powers of presidents: veto and decree Part II. Electoral Dynamics of Presidential Democracy: 5. Electoral dynamics: efficiency and inefficiency 6. Electoral rules and the party system 7. Electoral cycles and the party system Part III. Institutiona: Engineering: 8. Semi-presidentialism: the third alternative 9. Electoral cycles in semi-presidential regimes 10. Divided polities and collegial presidencies 11. Conclusions Appendices.
Article
This article examines the politics of what the French call 'cohabitation' in the period 1997-2002. It identifies two different ways in which cohabitation is studied. The first assumes that each of the three periods of cohabitation has occurred under a unique set of political circumstances. The second argues that there have been institutional similarities between each of the three periods of cohabitation and that fundamentally the political situation during each has been the same. The article provides evidence that supports both approaches, and concludes by suggesting that the second approach is the more useful.