How legislative democracy creates political parties Michael Koß, LMU Munich forthcoming in: Comparative Politics 51/4, 2019 Introduction Standard accounts on the rise of political parties emphasize the role of the electoral arena. Here, the nationalization of politics arguably led to a homogenization of voting behavior and, eventually, the emergence of coherent political parties. 1 From this perspective, the legislative arena appears as “chapter 2” of electoral democracy and is largely determined by party system properties. 2 The present article challenges this view and provides empirical evidence for Duverger’s claim that political parties can also emerge in the legislative arena. This argument acknowledges that the legislative arena often constitutes chapter one of democracy at large. 3 To capture the different dimensions of democracy, this paper introduces the term “legislative democracy” in which all legislation is subject to consent by assemblies. Legislative democracy often precedes electoral democracy as it may exist without universal suffrage, but not vice versa. Despite the temporal precedence of legislative democracy, scholars have primarily focused on the advent of electoral democracy. Previous research examined the choice of electoral systems, the enfranchisement of voters, and the arrival of the secret ballot. 4 In contrast, there exist hardly any analyses of the evolution of parties in legislatures apart from work done on the British House of Commons and the Congress of the United States. 5 This article aims to fill this gap by exploring the emergence of political parties in the lower chambers of two emerging European democracies, Sweden and France. The analysis focuses on the period between 1866 and 1958, i.e. the first wave of democratization. 6
32
Embed
How legislative democracy creates political parties
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
How legislative democracy creates political parties
Michael Koß, LMU Munich
forthcoming in: Comparative Politics 51/4, 2019
Introduction
Standard accounts on the rise of political parties emphasize the role of the electoral arena. Here,
the nationalization of politics arguably led to a homogenization of voting behavior and, eventually,
the emergence of coherent political parties.1 From this perspective, the legislative arena appears
as “chapter 2” of electoral democracy and is largely determined by party system properties.2 The
present article challenges this view and provides empirical evidence for Duverger’s claim that
political parties can also emerge in the legislative arena. This argument acknowledges that the
legislative arena often constitutes chapter one of democracy at large.3 To capture the different
dimensions of democracy, this paper introduces the term “legislative democracy” in which all
legislation is subject to consent by assemblies. Legislative democracy often precedes electoral
democracy as it may exist without universal suffrage, but not vice versa.
Despite the temporal precedence of legislative democracy, scholars have primarily focused on the
advent of electoral democracy. Previous research examined the choice of electoral systems, the
enfranchisement of voters, and the arrival of the secret ballot.4 In contrast, there exist hardly any
analyses of the evolution of parties in legislatures apart from work done on the British House of
Commons and the Congress of the United States.5 This article aims to fill this gap by exploring
the emergence of political parties in the lower chambers of two emerging European democracies,
Sweden and France. The analysis focuses on the period between 1866 and 1958, i.e. the first wave
of democratization.6
2
Legislative democracy creates political parties by means of two institutional mechanisms: control
over the plenary agenda in the plenary and powerful committees.7 The centralization of agenda
control in the hands of governments vertically differentiates the legislature and serves as an
incentive for legislators to exert an ex ante impact on legislation by means of voting discipline
maintained by political parties. Alternatively, powerful legislative committees horizontally
differentiate the legislature and allow legislators to exert an ex post impact on legislation. This
renders seats on powerful committees (and in particular committee chairs) legislative mega-seats,
i.e. desirable offices which allow for considerable impact on legislation.8 Given that parties serve
as the gatekeepers to seats on powerful committees, their establishment can be regarded as the
alternative path of legislative democracy creating political parties. These alternative paths pose the
question under which circumstances parties emerged as managers of government agenda control
or, alternatively, as gatekeepers of access to powerful committees.
In order to uncover the causal mechanism which allowed parties to emerge in legislatures, this
paper performs a comparative process-tracing analysis primarily based on transcript evidence that
has so far been largely overlooked: the committee reports and parliamentary proceedings
underlying procedural reforms. More specifically, the process-tracing analysis focuses on the 37
reforms that were debated on the plenary floor after the beginning of competitive elections in 1866
(Sweden) and 1871 (France), respectively. The Swedish and French legislatures are most similar
with respect to potential independent variables explaining the distribution of agenda control and
committee power, but most different regarding the ultimate outcome of procedural reform. In both
legislatures, a multi-party system evolved and coalition and minority governments were prevalent,
both of which were arguably conducive to the emergence of both decentralized agenda control and
powerful committees. Committees were indeed empowered in both legislatures; however, in 1958,
3
a procedural path change occurred in the French National Assembly during which agenda control
was centralized and committees were disempowered.
As the evidence presented here suggests, parties emerge as a response to an increasing quest for
procedural efficiency in the wake of the Industrial Revolution and the dawn of competitive
elections. Individual legislators preferred the creation of powerful committees which provided
them with a return for investing emerging party leaderships with the power to select committee
members, namely the prospect of mega-seats on committees. Since party membership became the
most important selection criterion for committee membership, the empowerment of committees
therefore caused the emergence of parties as gatekeepers of committee access.
In contrast, individual legislators were only willing to unilaterally surrender their inherited powers
to control the plenary agenda to party leaders if anti-system legislators could credibly threaten to
obstruct legislation. Unlike their pro-system counterparts, anti-system legislators have no interest
in cooperation with others. Rather, “an anti-system opposition abides a belief system that does not
share the values of the political order within which it operates” and aims to either break away from
the polity it operates in or to abolish democratic rule as such.9 Obstruction is here defined as the
exploitation of procedural loopholes in order to delay or derail legislation.10 Obstructive anti-
system legislators are willing to accept full-blown legislative deadlock. In this sense, maintaining
legislative democracy by means of a centralization of agenda control not only helps democratic
parties to thrive and survive, but also democracy at large.
Legislative organization and political parties
Scholarship on the advent of political parties mostly focuses on the electoral arena. Here, only
coherent political parties were able to fulfil voters’ demand for identifiable and coherent policy
4
platforms. In turn, candidates were increasingly elected on party tickets rather than individual
merits which allowed parties to impose discipline upon formerly autonomous legislators.11 In this
respect, party organization and party system properties such as fragmentation, patterns of
government alternation, or government formation are regarded as crucial for the evolution of
legislative organization.12
This article aims to show that this is not the whole story. Parties also emerge in the legislative
arena independent of and prior to electoral democracy. Representative institutions were originally
neither devised to foster democracy nor the emergence of political parties.13 Historically,
individual legislators dominated parliamentary procedure. This is why the “legislative state of
nature” is characterized by equal rights of all legislators.14 Two departures from the legislative
state of nature allow for the emergence of political parties in the legislative arena: vertical
differentiation, i.e. the centralization of agenda control and horizontal differentiation, i.e. the
establishment of powerful committees.
Agenda control refers to the ability to introduce, amend, and discuss legislation.15 These features
correspond to three dimensions of agenda control: timetable, positive, and negative control.
Timetable control refers to decisions about which proposals are debated. If governments or
majorities control the legislative timetable, they possess gatekeeping power. Positive agenda
control encompasses the power to amend legislation or prevent amendments, which is
accomplished through restrictive rules (which exclude certain kinds of amendments) or closed
rules (which forbid all amendments) as well as the power to make last amendments.16 Negative
agenda control comprises measures affecting the length of legislative debates such as closure
procedures, which terminate such debates.
5
Agenda control can be defined as centralized if governments or majorities enjoy privileges in two
of the three dimensions (timetable, positive, and negative agenda control). The centralization of
agenda control is an institutional prerequisite for the dominance of political parties in the
legislative arena because serves as an incentive for the cohesion of governing parties and
opposition parties alike.17 This cohesion ensures that governments include the policy preferences
of their legislators in an ex ante fashion: Since governments cannot rely on compromises with
disciplined opposition parties, they have to ensure their proposals reflect the views of rank-and-
file legislators to get passed.
The second institutional mechanism allowing parties to dominate legislative organization are
powerful committees. Committees provide coalition partners without control of executive
departments with the expertise necessary to control the actions of the respective minister and to
either alter or stop proposals which deviate from their preferences.18 In order to enable legislators
to do so, three features of committees are essential: First, committees need to be able to rewrite
bill proposals, which not only allows them to change government proposals but also ensures that
it is the amended committee version of the proposal which becomes the basis of the plenary debate.
Second, permanence, which allows committee members to acquire the expertise which
distinguishes them from their ordinary peers not sitting on committees. Third, committees need to
be congruent with executive departments which makes it more likely that committees are able to
oversee the activities of particular ministers.
Committees are powerful if they possess any combination of two of their three core features
(permanence, rewrite authority, and congruence with executive departments). Powerful
committees allow individual legislators to affect policies in an ex post fashion because they ensure
that proposals which were controversial in the cabinet stage receive more scrutiny and
6
amendments in the committee stage.19 Given that committee seats allow for enhanced ex post
impact on legislation, powerful committees have the same effect as a centralization of agenda
control: they foster the emergence of political parties as the central actors in legislatures. Political
parties emerge as the only feasible gatekeepers to committee seats, which in turn serves as an
incentive for individual legislators to keep to the party line in order to ensure promotion to mega-
seats on powerful committees. Dissenters lose their committee seats more often than those who
hold to the party line.20
To summarize, control over the plenary agenda and powerful committees are alternative
institutional mechanisms which ensure the emergence of parties in the legislative arena. In the case
of centralized agenda control, this is ensured by means of party discipline allowing individual
legislators to affect policies in an ex ante fashion. Governments are forced to include individual
legislators’ views in their proposals because there is no chance for policy deals with opposition
legislators in committee. In the case of powerful committees, parties become the gatekeepers of
access to sought-after committee seats which allow individual legislators to amend government
proposals and influence policies in an ex post fashion.
The origins of centralized agenda control and powerful committees
The existence of two institutional mechanisms ensuring the emergence of parties in legislatures
raises the question of how to explain the respective departures from the legislative state of nature.
As for control over the legislative agenda, previous research identified the fragmentation of party
systems and alternation in government as the major causes for a centralization. The smaller the
number of parties represented in a legislature and the more regular (coalitions of) parties alternate
in power, the more likely a centralization of agenda control arguably becomes.21 Such correlations
between the fragmentation of party systems, patterns of government alternation, and the
7
centralization of agenda control exist. However, from the perspective of causality, this raises the
question why, even in legislative democracies with emerging two-party systems and regular
alternation in government, procedurally equal junior legislators would accept to invest party
leaders with far-reaching powers over the plenary agenda.
The origins of powerful committee systems have received less scholarly attention, at least beyond
the U.S. Congress.22 With respect to parliamentary systems, André, Depauw, and Martin present
the most elaborate analysis. Based on the notion that committees are monitoring devices of
coalition partners, they argue that the advent of coalition government empowered committees that
helped coalition partners keep tabs on each other.23 Similarly, Strøm notes a correlation of
powerful committees and minority governments, which implies that committees are the arenas
where governments negotiate policy deals with parliamentary majorities.24 From the perspective
of causality, however, this also raises questions, most notably how to explain the advent of
powerful committee prior to the institutionalization of parliamentary government. An answer to
the questions raised here requires a causal mechanism to be corroborated in a process-tracing
analysis of procedural reforms.
Methodological underpinnings
A process-tracing analysis is a “procedure for identifying steps in a causal process leading to the
outcome of a given dependent variable of a particular case in a particular historical context“.25
This focus on causal processes renders it appropriate to identify the logical steps that led to the
emergence of parties in the legislative arena. More specifically, each logical step requires a causal
process observation as “an insight or piece of data that provides information about context or
mechanism and contributes a different kind of leverage in causal inference”.26 If causal-process
observations can be contextualized in a particular case (here: of procedural reform), they count as
8
evidence for the existence of a causal mechanism.27 Causal mechanisms are “portable concepts
that explain how and why a hypothesized cause, in a given context, contributes to a particular
outcome”.28
The evidence analyzed to identify causal process observations will be the parliamentary
documentation relating to particular reform attempts, such as the reports of procedural committees
and parliamentary records of reform debates. Such transcript evidence “offers a unique opportunity
to assess actors’ beliefs”.29 To be sure, reasons forwarded by legislators do not necessarily equal
causes explaining procedural changes. Additionally, public speeches are often regarded as
problematic evidence since actors tend to conceal their true intentions.30 Indeed, it is often
advisable to take legislators’ explanations for their actions with a grain of salt. However, there are
two reasons why parliamentary documents suffer from less bias than other transcript evidence.
First, the parliamentary floor is an arena of interaction, which implies that actors concealing true
intentions and true causes will most likely be exposed by their peers. Second, parliamentary
transcript evidence can be triangulated with contemporary and current scholarship on parties and
legislatures in order to trace back reform processes and their causal origins.
In which legislatures and during which time period shall transcript evidence on which procedural
reforms be analyzed to trace the emergence of parties? The selection of cases requires that the
spatial, temporal, and substantive boundaries of the analysis are determined.31 The cases to be
analyzed are legislatures in emerging parliamentary democracies which are most similar with
respect to potential independent variables but still display most different outcomes of procedural
reform. As regards spatial boundaries, the analysis focuses on the Second Chamber of the Swedish
Riksdag (Andra kammaren) and the French Chamber of Deputies (Chambre des Deputés), which
9
became the National Assembly (Assemblée nationale) in 1944. The procedural evolution of these
legislatures has hardly been analyzed.32
The legislatures and the emerging parliamentary party systems of Sweden and France showed
several similarities: Both were directly elected lower chambers in two-chamber systems where
lower and upper chambers possessed equal legislative powers.33 Another similarity lies in the
multi-party systems which evolved in both the Swedish and French legislatures. By the 1870s, two
embryonic parties had emerged in the Riksdag’s Second Chamber, a dominant conservative
Ruralist Party and a liberal party. The Social Democrats entered the legislature in 1895. However,
legislators could only meaningfully be associated with party labels since 1911.34 After 1911, a
fluid five-party system emerged when a Farmers’ Party was created in 1917 and a faction of the
Social Democrats declared itself the Communist Party in 1921. This five-party system stabilized
after 1934 and was henceforth dominated by the Social Democrats. Parliamentary parties in France
were similarly factionalized as their Swedish counterparts. After 1871, a Republican and a
Conservative camp emerged with several sub-groups. Socialists were first represented in 1893 and
also experienced internal struggles, which also led to the creation of a Communist Party in the
wake of a split in 1921. In 1901, the Radicals emerged as the pivotal parliamentary party and
remained so for the rest of the Third Republic. After 1944, the parliamentary parties of the Fourth
Republic consisted of three mutually opposing groups: laical Radicals and clerical Christian
Democrats, Socialists and Communists, and Gaullists who opposed all other parties.35
Both legislatures are also remarkably similar with respect to potential independent variables. First,
in both of them, Communist anti-system parties existed which may have caused a centralization
of agenda control. Just as well, the emerging multi-party system may have prevented the
centralization of agenda control. Likewise, parties in both countries formed coalition or minority
10
governments that were arguably conducive to powerful committees once the parliamentary
responsibility of governments was established (in France in 1876 and in Sweden in 1919).
As for the temporal boundaries of case selection, the period of investigation starts with the advent
of elected legislatures, i.e. assemblies which needed to approve or consent to all legislation. This
was the case in Sweden in 1866, when the former four-chamber Estates Riksdag became an elected
body. The Second Chamber was directly elected through majority rule, initially by only about 20
per cent of the male population. When proportional representation was introduced in 1909, the
suffrage was extended to about 40 per cent of the male population. Since 1921, members of the
Riksdag’s Second Chamber have been elected by universal suffrage.36 In France, the analysis starts
in 1876, when the Chamber of Deputies of the Third Republic was elected for the first time by
universal male suffrage. In 1919 and 1945, variants of proportional representation were
introduced, but they were abolished both times (in 1927 and 1958, respectively) in favor of the
original majoritarian system with two ballots introduced in 1876.37 The 1958 transition towards
the French Fifth Republic marks the end of the period of investigation, since most different
procedural outcomes – on the one hand, powerful committees (Riksdag), and on the other,
centralized agenda control (National Assembly) – persist to this day.38
As regards the substantive boundaries of the case selection, any proposal aiming to (de)centralize
control over the legislative agenda and either strengthen or weaken legislative committees which
became subject to a plenary debate will be taken into consideration. This results in a total of 37
procedural reform attempts in the Swedish and French legislatures. Figures 1 and 2 provide further
information on these reforms. Each of the 37 reform attempts is represented by a diamond on the
respective line in the figures. There were 15 attempts in the Swedish legislature and 22 in the
French legislature. 16 of the 37 reform attempts concerned agenda control and 21 committee
11
powers. Further information on all these reform attempts and the sources for evidence on these
reforms and the underlying parliamentary procedures and institutions can be found in the online
appendix.
Figure 1: The Evolution of legislative organization in the Riksdag, 1866–1958
Note: Dots indicate reform attempts.
Initially, all three dimensions of agenda control (timetable, positive, and negative) were
decentralized in both legislatures, preventing parties from dominating plenary proceedings. In the
Riksdag as well as in the Chamber of Deputies, only one feature of committee power was present
(permanence in Sweden and rewrite in France), which meant that committees were initially weak.
A mere five of the 37 attempts changed the distribution of power regarding committees or agenda
control in both legislatures. This is visualized by the solid and dotted lines (agenda control and
committee power, respectively) of figures 1 and 2 going up or down. First, the Chamber of
Deputies’ committees were empowered in 1902 when committees (which, as mentioned, already
possessed rewrite authority) became permanent. Second, the permanent Swedish committees were
empowered in 1908 when they received rewrite authority. Third, the formal disempowerment of
committees in 1935 remained ineffective.39 The fourth and fifth changes occurred simultaneously
0
1
2
3
1875 1900 1925 1950
Agenda control (dots indicate reform attempts) Committee power (dots indicate reform attempts)
No
gove
rnen
t con
trol
Gov
erne
nt c
ontro
l
Cen
traliz
ed d
imen
sions
12
and were the most fundamental ones as figure 2 suggests. In 1958, the newly adopted constitution
of the French Fifth Republic invested governments with timetable, positive, and negative agenda
control. Additionally, committees lost their authority to rewrite proposals and were accordingly
disempowered.
Figure 2: The Evolution of legislative organization in the Chamber of Deputies / National
Assembly, 1876–1958
Note: Dots indicate reform attempts. The grey area depicts the period of breakdown of legislative democracy under
German occupation.
The 1958 reform in the French National Assembly meant that despite their many similar features,
the legislatures in both countries nonetheless differ with respect to the ultimate outcome of
procedural reform. This begs the question whether party system properties caused these changes
or whether changes to legislative democracy created political parties. The following three sections
aim to provide transcript evidence to answer this question.
0
1
2
3
1875 1900 1925 1950Agenda control (diamonds indicate reform attempts) Committee power (diamonds indicate reform attempts)
No
gove
rnen
t con
trol
Gov
erne
nt c
ontro
l
Cen
traliz
ed d
imen
sion
s
Ger
man
occ
upat
ion
13
The quest for efficiency and the ascendancy of parties
The evidence from all 37 reform attempts clearly suggests that the primary cause of procedural
reform was not related to political parties. With the exception of reforms codifying legislative
states of nature at the outset of the Third and Fourth French Republic (numbers 16 and 32 in the
online appendix) and one proposal rejected on the floor of the Riksdag without discussion (number
1 in the online appendix), all reform proposals which reached the plenary stage referred to the
increased legislative workload or the need to organize legislative procedures more efficiently.
References for all these observations are presented in the online appendix. Only the push for
efficiency that arose from legislative democracy allowed political parties to emerge in the
legislative arena.
The 1908 Riksdag proposal empowering committees, the most important reform in Sweden,
exemplarily illustrates the relation between growing legislative workload and the need for
procedural reform. The proposal explicitly referred to recent economic and societal changes that
prompted a considerable increase in the number of bill proposals, especially ones of “financial and
social nature”.40 The report also provided numbers for proposals, amendments, session schedules,
etc. This suggests that procedural reforms were explicitly devised to cope with the growing state
activity in the wake of the Industrial Revolution relating to the regulation of new policy areas,
most notably economic and social policy.41 Additionally, procedural reforms were means to
increase parliamentary efficiency. In Sweden, a more “rational” organization of work in
parliament was supposed to help complete business more quickly.42 Arguably, the inclusion of
minorities in the committee stage would ideally help to effectively multiply the plenary arena and
avoid long plenary debates.43
In the French Chamber, legislative procedures were regarded as too slow, especially in light of the
fact that committee membership, which was selected by lot, did not mirror the composition of the
14
floor which prevented an effective multiplication of the plenary. This arguably led to a growing
“incoherence” in legislative procedure.44 Consequently, as in Sweden, reform proposals called for
a more “rational” procedure.45
The democratization of committee membership and the ascendancy of parties
The transcript evidence of the 21 reform attempts relating to committee powers suggests that the
growing demand for a democratization of access to committees caused parties to become
gatekeepers of said access. As expected, this allowed party leaders to become the central actors in
legislatures. In both the Swedish and the French legislatures, committee seats could be allocated
arbitrarily prior to the empowerment of committees. In the Riksdag, a strong norm dictated that
senior legislators would always receive committee seats independent of partisan majorities.46
Against the background of heightened political tensions, legislators increasingly warned that
majority rule could be “abused” by governmental majorities and prevent minorities from being
represented on committees.47 Given that the Riksdag’s permanent committees coordinated the
passage of legislation between both chambers, committee membership was highly desirable for
Swedish legislators. Against this background, the seniority-based allocation mechanism of
committee seats caused increasing discontent. Even though the number of permanent committee
seats grew from 48 to 72 between 1866 and 1945, Swedish legislators were on average only
appointed to a permanent committee during their third term in the Riksdag’s Second Chamber.48
As a consequence, individual legislators called for an extension of the committee system since
mere participation in plenary debates was regarded as “boring”.49 These growing requests are
ideal-typically exemplified by the speech that Albin Ström, a Social Democrat, held in 1932. Ström
likened the internal organization of the Riksdag to the employment sector and called legislators
not represented in permanent committees “jobless”.50 According to Ström, legislators without
15
committee seats were handicapped since they could hardly contribute to the debate in the plenary,
which was based on committee proceedings. In 1948, a cross-party motion called for a reform of
the committee system ensuring the “employment” of all legislators.51
In the French Chamber of Deputies, the arbitrary allocation of committee seats raised similar
discontent. Here, majority elections of committee members in bureaus chosen by lot allowed
senior legislators membership in multiple committees. Even though the maximum number of
committee seats that could be held by the same legislators was set to two in 1876, the Speaker
soon admitted that he was powerless to enforce this rule.52 In 1894, the independent Marcel Habert
criticized that “it was always the same [parliamentarians] who get elected [to committees]”.53 As
in Sweden, committee seats were regarded as highly desirable, but more as steps towards the
cabinet rather than instruments of inter-chamber coordination. Accordingly, French legislators
criticized the fact that “large parts” of their peers were “excluded” from committees.54 Against this
background, bureau elections of committee members were seen as producing a “salon of the
refused”.55 Similar to their Swedish peers, French legislators also called for more “occupation” in
committees.56
Given the increasing demand for seats and the growing criticism of the existing allocation
mechanisms for committee seats, political parties came to be regarded as mediators which could
best ensure fair representation of legislators in committees. By 1908, even the Swedish justice
minister Albert Petersson, despite being an independent not holding a Riksdag seat, had to admit
that the distribution of committee seats through party groups was “particularly appropriate” given
that it would ensure the “impartial evaluation” of proposals and a “bridging of party oppositions”
in committee.57 Before 1908, senior legislators had denied parties the role of gatekeepers of
committee seats which would allegedly foster “partisan conflicts” preventing the best decisions on
16
policy proposals.58 Accordingly, the 1908 reform turned parties into gatekeepers of committee
access and ended the dominance of the seniority principle. Tellingly, justice minister Petersson
also came to affiliate with the Conservatives after the 1908 reform.59 Even though individual
legislators could formally present lists of names to be approved for committee membership, party
groups always created these lists centrally. Parties’ positions as gatekeepers henceforth allowed
party leaders to discipline dissenting legislators without any formal sanctioning mechanism.60
In France, the reform proposal which preceded the empowerment of committees in 1902 also
envisaged parties as gatekeepers in order to ensure that “all deputies who wish to work are able to
find a committee employment corresponding with their competences without running the risk of
being evicted completely by the hazard of bureau elections”.61 Due to the ongoing suspicion vis-
à-vis parties among conservative legislators, it was not until 1910 that parties finally became the
gatekeepers of committee access. By this time, even conservatives like Edouard Aynard, chair of
the procedural committee, called parties the “very essence of the parliamentary regime”.62 Since
1915, registration with party groups was compulsory to ensure that all legislators received a
committee seat.63
The evidence presented here suggests that parties became central actors in parliaments since they
were able to offer legislators a more predictable career path than the pre-democratic seniority-
based status quo ante. In contrast, there is no evidence for a causal relation between the dominant
mode of government formation and committee power. In Sweden, neither parliamentary
government nor meaningful party labels existed when committees were empowered in 1908. In
the French parliamentary regime, committee membership was indeed regarded as a step towards
the cabinet, but there is no evidence that parties used committees to keep tabs on each other. In
both legislatures, parties hardly existed when committees were empowered. Rather, parties
17
eventually emerged as central actors in parliaments as a response to bottom-up demand for
committee representation, which in turn allowed them to emerge as the fairest gatekeepers of
access to newly empowered committees.
The centralization of agenda control and the ascendancy of parties
The transcript evidence of all 16 reform attempts relating to agenda control suggests that anti-
system obstruction explains the effective delegation of agenda-setting powers to governments and,
eventually, political parties. The mere existence of an anti-system actor did not suffice to trigger
successful reforms. Irrespective of the Communist SKP, a Swedish commission of inquiry even
rejected a modest centralization of timetable control since “in the Swedish Riksdag, it has never
occurred that discussions were extended with the aim of obstruction”.64 This suggests that an anti-
system party is only causally relevant for procedural reforms in conjunction with legislative
obstruction. Like in Sweden, obstruction was initially absent in the French Chamber even after the
advent of the Communist PCF. Nonetheless, the presence of the more orthodox French
Communists caused unrest, which is illustrated by the large number of procedural reform
proposals after 1920 (see reforms 27–36 in the online appendix). Given that “attacks directed
against representative democracy”65 remained nonetheless confined to the extra-parliamentary
arena, agenda control continued to be decentralized.
This picture changed when the Communists left de Gaulle’s post-war all-party government in
1947. From then on, they adopted obstructionist tactics and “exploited procedural loopholes with
persistence and ingenuity”.66 As a response, senior legislators in the National Assembly agreed on
reforms aiming to centralize agenda control: As early as 1948, the procedural committee warned
that Communist obstruction would otherwise “discredit parliamentarism”.67 Similarly, a 1952
report aimed to counter “the sabotage of legislation”.68 However, senior legislators failed to
18
impose discipline on their colleagues. This revealed the increasing weakness of parties which had
initially become central actors in the legislative arena in the wake of the empowerment of
committees. In 1946, even a proposal to abolish spontaneous amendments during plenary sittings
was rejected on the floor. Even though the powers of the steering body grew in 1952 and 1954
(see reforms 28–29 in the online appendix), the plenary remained able to overturn all central
decisions regarding the legislative timetable.69 As a consequence, governments linked decisions
on the timetable to confidence votes – and stepped down in case of defeats, which explains cabinet
instability in the Fourth Republic.
Only when the threat of a military coup against what Conservatives perceived as “the permanent
treason of the communist party”70 during the Algerian war grew into a critical juncture in 1958, a
majority of legislators surrendered their influence on procedural reform. When the National
Assembly granted extraordinary powers (most notably, to draft a new constitution) to the new de
Gaulle government on 1 June 1958, many non-Gaullist legislators justified their decision by
invoking Communist obstruction. The Socialist Maurice Deuxonne declared that he would deviate
from his party line for the first time and vote for de Gaulle since the previous legislative period
had been wasted on a “class struggle”.71 According to fellow Socialist Jean Le Bail, the Fourth
Republic suffered from the “poisoned gifts” of proportional representation and the PCF.72
Similarly, even de Gaulle’s political opponent, the former Radical Prime Minister Pierre Mendès-
France, argued that it was not the parliamentary regime but rather its abuse which caused the
failure of the Fourth Republic.73 The Christian Democrat François de Menthon summarized the
discussion when he stated that several deputies were planning to vote for de Gaulle because they
feared the PCF. De Menthon himself, however, announced his intention to vote against de Gaulle
because he reasoned that the full powers for the general would strengthen the PCF even further.74
19
Once de Gaulle had taken office, the procedural part of the constitution was primarily drafted by
Michel Debré, a close aide of the Prime Minister who explicitly aspired to reduce the influence of
political parties in general and the PCF in particular. Debré thought of the PCF as “a tool of the
total claim to power” directed by Moscow whose aim was to destroy the French Republic.75 Debré
aimed to privilege governments procedurally in the constitution because he believed that electoral
reform alone would not necessarily produce majorities given the strength of the PCF.76
The explanatory power of anti-system obstruction for the centralization of agenda control is
illustrated by the fact that senior legislators also approved the constitutional reform. De Gaulle’s
cabinet accepted virtually all of the procedural changes that were envisaged by Debré. His
ministers were well aware that strict measures were necessary to overcome the legislative deadlock
that characterized the previous sessions.77 Similarly, senior legislators represented in the
Consultative Constitutional Council involved in the drafting of the constitution made hardly any
objections either, especially with respect to the centralization of timetable and negative agenda
control.78 The Consultative Council was only opposed to the centralization of positive agenda
control by means of the block vote and recommended to leave the right to amend legislation
unrestricted.79 Debré vetoed this recommendation because the aim of the block vote was to counter
“obstructionist methods”, as one of his advisors put it.80
The evidence presented here suggests that anti-system obstruction was causally crucial for the
centralization of agenda control in the National Assembly. The high fragmentation of the French
party system did not prevent this reform. On the contrary, Debré advocated the centralization of
agenda control as a means to contain the electorally powerful PCF and to overcome factional
disputes.81 As long as the PCF did not obstruct, however, a majority of legislators refused all
substantial reforms of agenda control. Only when a critical juncture emerged, this majority became
20
willing to surrender its procedural privileges. Perhaps the biggest paradox in the 1958 transition
towards the Fifth Republic was that Debré, one of the biggest critics of party rule, ensured its
continuity beyond a period of sustained anti-system obstruction.
A causal mechanism explaining the ascendancy of political parties in the legislative arena
There are two alternative paths which lead to the emergence of political parties in legislatures
(figure 3).82 This article has primarily focused on the first four steps up to the empowerment of
committees in step 5a or the centralization of agenda control in step 5b. How parties dominate the
legislative arena once powerful committees exist or agenda control is centralized is well-
established in the literature.83 The arrows in figure 3 depict necessary conditions. Steps 1–6a and
1–6b, respectively, are jointly sufficient and individually necessary for the emergence of parties
in legislatures. From this follows that legislators always have a choice and that there is no
determinism leading from the first to the last step in the causal mechanism.84
According to the process-tracing analysis of 37 procedural reform attempts in the Swedish and
French legislatures, individual legislators themselves triggered their own decline as equal actors
in legislatures by requesting ever more plenary time as a response to the emergence of new policy
areas in the wake of the Industrial Revolution. Confronted with the need to regulate a growing
number of policy areas and constrained by the democratic requirement of assembly consent to all
legislation, individual legislators resorted to procedural reforms increasing legislative efficiency
(step 1).85 In this sense, legislative democracy has an inherent dynamic to create hierarchies
between legislators and, eventually, political parties. Absent the need to increase legislative
efficiency, reforms of agenda control and committee power remained ineffective.
21
Figure 3: The advent of parties as central actors in the legislative arena
(1) Pressure to increase efficiency
¢ no
Non-reform
¢ yes
(2) Anti-system obstruction
¢ no
(3a) Committees differentiate the plenary horizontally
¢
(4a) Parties become fair gatekeepers of committee seats
¢
(5a) Creation of powerful committees
¢
(6a) Incentive for legislators to affiliate with parties
¢
(7) Parties central actors in legislative arena
¢ yes
(3b) Critical juncture
¢
(4b) Legislators willing to surrender privileges
¢
(5b) Centralization of agenda control
¢
(6b) Parties organize use of agenda control
¢
Arrows (¢) depict necessary conditions; “no” and “yes” refer to the absence or presence of causal factors.
From the perspective of individual legislators as the historical bearers of procedural power in
legislatures, the creation of powerful committees is the more attractive option to increase
legislative efficiency (see steps 3a–6a in figure 3). Powerful committees allow to horizontally
differentiate legislatures by creating several arenas in which legislative proposals can be dealt
with. Such a horizontal differentiation creates more career opportunities than a vertical one by
means of a centralization of agenda control. Especially the chairs on powerful committees, but
also the position as reporter of important committee proposals in the plenary constitute desirable
mega-seats to which no equivalent exists under centralized agenda control.86 Against this
background, legislators propose to delegate substantial law-making powers to committees (step
3a). This delegation occurs if legislators also accept parties as fair gatekeepers of access to
committees (step 4a). The pre-democratic seniority principle fails this fairness test. Parties are
more desirable as gatekeepers from the perspective of individual legislators because party
membership can be easier acquired than expertise and/or experience. Once powerful committees
22
are created, legislators have an incentive to affiliate with parties which then become the central
actors in the legislative arena (see steps 6a and 7).
Only the intervention of an extraordinary event can reasonably explain why legislators forfeit
career opportunities and accept the centralization of agenda control. Anti-system obstruction is
such an extraordinary event which is able to convince individual legislators to forfeit both their
ancient procedural privileges and the career prospects of powerful committees (step 2). Anti-
system legislators have no scruples to engage in unlimited obstruction against all legislation and
provoke a breakdown of legislative democracy. If such anti-system obstruction constitutes a
critical juncture (step 3b), it becomes possible that legislators accept the procedural empowerment
of governments by means of a centralization of agenda control (step 4b). Political Parties are the
only feasible bearers of agenda control able to overcome an anti-system threat on behalf of
governments by means of coordinated voting behavior. Similar to the creation of powerful
committees, the centralization of agenda control accordingly leads to the advent of political parties
as central actors in the legislative arena which ensure the cohesion of coalition partners and the
opposition (see steps 6b and 7).
Conclusion
A comparative process-tracing analysis of all 37 procedural (non-)reforms of committee power
and agenda control in the Swedish and French legislatures during the 1866–1958 period suggests
that political parties emerged from the frictions of legislative democracy. Legislators growing
activity in the wake of the Industrial Revolution and the advent of electoral democracy triggered
a quest for procedural efficiency during which parties became the central actors in legislatures
through, above all, two institutional mechanisms: a horizontal differentiation of powerful
committees or a vertical differentiation through a centralization of agenda control. The former
23
provides individual legislators who covet mega-seats with an incentive to follow the party line.
The latter allows party leaders to impose discipline on legislators.
The evidence presented here suggests that legislative democracy and political parties co-evolve.
Rather than the effective number of parties, their alternation in government, or patterns of
government formation that parties engage in, the interactions of (emerging) senior and junior
legislators explain how parties came to dominate legislative procedure. During these interactions,
the occurrence of anti-system obstruction accounts for the institutional mechanism through which
this dominance is established because it allows for the (exceptional) vertical differentiation of
legislatures. Absent anti-system obstruction, legislators’ demand for committee seats triggers the
more common horizontal differentiation by means of an empowerment of committees. The
resulting co-evolution of legislative democracy and parties explains why there is also evidence
that the rules of legislative organization explain party system properties.87
Acknowledgements
This paper has been presented in a panel on the evolution of legislative institutions at the 2017
ECPR General Conference in Oslo. The author wishes to thank all participants and the two
anonymous referees. Generous funding for this research was provided by the Volkswagen
Foundation (grant number 88252).
1 Daniele Caramani, The Nationalization of Politics: The Formation of National Electorates and
Party Systems in Western Europe (Cambridge: Cambridge University Press, 2004).
2 Royce Carroll, Gary W. Cox, and Mónica Pachón, “How Parties Create Electoral Democracy,