Content uploaded by Iñaki Sagarzazu
Author content
All content in this area was uploaded by Iñaki Sagarzazu on Jun 29, 2018
Content may be subject to copyright.
European Union Politics
14(3) 388–407
!The Author(s) 2013
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1465116512472938
eup.sagepub.com
Article
Ideological congruency and
decision-making speed:
The effect of partisanship
across European Union
institutions
Heike Klu
¨ver
University of Konstanz, Germany
In
˜aki Sagarzazu
University of Glasgow, UK
Abstract
How does ideological congruency affect the speed of legislative decision-making in the
European Union? Despite the crucial importance of actor preferences, the effect of
partisan alignments and ideological composition of the European institutions has largely
been neglected. However, we argue that the ideological congruence between legislative
bodies has an important effect on the duration of policy-making. We test our theoret-
ical expectations based on a large new dataset on decision-making speed in the
European Union using event history analysis. The findings confirm our theoretical
claim indicating that the ideological distance between the European institutions slows
down policy-making which has important implications for the problem-solving capacity
of political systems more generally.
Keywords
Decision-making, duration, European Union, ideology, preferences, speed
Introduction
Why does decision-making speed vary across proposals?
1
Rather than analyzing
the duration of policy-making, legislative studies largely focus their attention on
whether a proposal gets approved at all. Drawing on theories of cartelized
Corresponding author:
In
˜aki Sagarzazu, School of Social and Political Sciences, University of Glasgow, Adam Smith Building, Glasgow
G12 8RT, UK.
Email: inaki.sagarzazu@glasgow.ac.uk
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
legislatures (e.g. Cox and McCubbins, 1993), unified or divided government (e.g.
Howell et al., 2000) and of legislators’ ideological positions (e.g. Calvo and
Sagarzazu, 2011), legislative scholars have attempted to explain the approval of
policy proposals. However, the duration of the policy-making process has received
only little attention which is surprising, as an essential quality of modern policy-
making is decision-making speed. Policy problems require increasingly fast legis-
lative reactions so that a crucial feature of the problem-solving capacity of political
systems is the ability to quickly respond to new policy problems. Several studies
have accordingly pointed out the important role of time in politics (see e.g. Do
¨ring,
1995; Goetz, 2009).
The few existing studies of decision-making speed have typically analyzed
policy-making in the European Union (EU) which involves the European
Commission (EC), the European Parliament (EP) and the Council of the EU
(Golub, 1999, 2002, 2007, 2008; Golub and Steunenberg, 2007; Ko
¨nig, 2007,
2008; Rasmussen and Toshkov, 2010; Schulz and Ko
¨nig, 2000). Schulz and
Ko
¨nig (2000) as well as Golub (1999, 2002, 2007) and Golub and Steunenberg
(2007) find that decision-making speed in the EU slows down if the EP is formally
involved in the policy-making process. Similarly, these studies agree that qualified
majority voting in the Council of Ministers by contrast shortens the duration of
European policy-making.
While the effect of institutional characteristics on decision-making speed has
received considerable attention, the impact of actor preferences has only been dis-
cussed with regard to the ideological diversity within one particular legislative
body. Golub (1999, 2002, 2007) and Golub and Steunenberg (2007) analyze the
effect of ideological diversity in the Council by including a dummy variable to
capture the effect of Margaret Thatcher as an example of an extreme preference
outlier who delays the decision-making process. Schulz and Ko
¨nig (2000) infer the
distribution of preferences of member states in the Council drawing on the issue
area. They argue that member states have relatively homogeneous preferences in
internal market policy, agricultural policy, competition policy and external trade
whereas their preferences diverge in other policy areas. Ko
¨nig (2007) further
improves the operationalization of member state preferences by drawing on
national party manifestos to estimate policy positions in four policy areas and
on a pro/anti European integration dimension.
Previous studies of policy-making duration in the EU have considerably
enhanced our understanding of legislative decision-making speed by investigating
the effects of institutional characteristics and ideological congruence within one
particular legislative body. However, the effect of ideological diversity between
different legislative bodies on the timing of policy-making processes has not been
investigated with the exception of Toshkov and Rasmussen (2012) and Dru
¨ner
et al. (2012) who only focus on a small number of bills.
2
This is surprising as
we know from the literature that the distance between decision-making bodies
has a crucial impact on policy-making outcomes (Tsebelis, 1995, 1999, 2002).
We therefore argue that in settings with multiple institutional actors involved in
Klu
¨ver and Sagarzazu 389
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
the legislative process, the ideological distance between the legislative bodies that
hold the most extreme policy positions crucially affects the speed with which legis-
lation is approved. We empirically test our theoretical claim drawing on a novel
dataset that includes all legislative proposals introduced in the EU between 1979
and 2010.
Studying the effect of ideological diversity on decision-making speed in the EU
contributes to the ongoing debate about the nature of political conflict in the EU.
While national political systems are dominated by the fierce competition between
political parties on the traditional left–right dimension, the nature of political con-
flict in the EU has generated a vibrant debate. First, scholars examining the emer-
gence of the EU largely identify a conflict that revolves about the degree of
European integration (e.g. Moravcsik, 1998; Sandholtz and Stone Sweet, 1998).
Second, another strand of research has empirically evaluated the policy space at the
European level and concluded that it is characterized by a unidimensional left–
right space (Gabel and Hix, 2002; Hix et al., 2005; Kreppel and Tsebelis, 1999;
Tsebelis and Garrett, 2000). Third, a number of studies have contested these con-
clusions and argued that political conflict in the EU is by contrast characterized by
a two-dimensional space consisting of a left–right and a pro-anti European inte-
gration dimension (McElroy and Benoit, 2007). In addition, scholars examining the
internal functioning of the European institutions have debated the role of ideology
versus nationality. For instance, even though McElroy and Benoit (2010) have
shown that national parties join EP party groups due to ideological orientations,
other studies have shown that voting behavior of MEPs is ultimately determined by
national parties rather than their EP party groups (Hix, 2002, 2004). Similarly,
while Hagemann and Høyland (2008) have shown that ideological affiliations influ-
ence coalition formation in the Council, others have argued that decisions are
driven by national and regional, rather than partisan cleavages (e.g. Mattila and
Lane, 2001). The findings of the present study therefore also have important impli-
cations for the ongoing debate about the nature of political conflict in the EU.
This article shows that the ideological composition of the European institutions
has a crucial effect on decision-making speed. If the EC, the Council and the EP
have similar preferences, proposals are quickly adopted as there is no need for
extensive bargaining. However, if they have diverging preferences, the decision-
making process is delayed by bargaining between the European institutions which
try to amend the proposal in their favor. Hence, ideological differences between the
European institutions slow down the policy-making process.
Ideological congruency and policy-making speed
In this section, we present the central theoretical argument that explains decision-
making speed with the ideological congruence of legislative bodies. As the theoretical
argument not only applies to the EU, but to all political systems in which more than
two legislative bodies are involved in policy-making, we present the argument in
general terms. The first step in explaining the duration of policy-making is the
390 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
identification of the decisive set of actors. Tsebelis (1995, 2002) suggested that pol-
itical actors matter if they classify as a ‘‘veto player’’. A veto player is an ‘‘individual
or collective actor whose agreement is required for a policy decision’’ (Tsebelis, 1995:
293). Even though the number and design of political institutions vary considerably
across different political systems, this general definition allows to determine the
decisive actors irrespective of the institutional setting in which they are embedded.
Whereas Tsebelis (1995, 2002) however largely focused on explaining policy change
or stability, the aim of this study is to explain the speed with which legislation is
approved.
The veto player definition provided by Tsebelis is perfectly reasonable for the
analysis of policy change, but political actors can prolong the duration of policy-
making without having veto power. For instance, Tsebelis and Money (1997) show
that upper chambers without veto power are able to delay legislation until the next
election in an effort to block the adoption of a proposal until new parliamentary
majorities are created. Similarly, Kardasheva (2009) demonstrates that the EP can
slow down the policy-making process in the EU even if it does not have any veto
power over a legislative proposal. It is therefore necessary to extend the definition
to cover all political actors that can affect the timeliness of decision-making. The
crucial actors in policy-making are accordingly all legislative bodies whose involve-
ment is mandatory for the adoption of legislation. For reasons of analytical par-
simony, we treat these legislative bodies as unitary actors.
3
We assume that legislative bodies are rational, goal-oriented and purposeful
collective actors (Downs, 1957). We assume that decision-makers carefully weigh
the costs and benefits of political alternatives and select those conferring maximum
utility. We assume that legislative bodies have fixed exogenous preferences over
policy outcomes. They evaluate the utility of a policy proposal based on its prox-
imity to their own ideal point (Tsebelis, 1995, 2002). The closer a policy proposal to
their preferred position, the higher the utility they attach to this legislative initia-
tive. Rational legislative bodies therefore aim at adopting legislation that is as close
as possible to their own preferred position.
Policy-making however involves several legislative bodies which pursue their own
policy preferences. For instance, most of the legislation adopted in Germany requires
the consent of both chambers, the Bundestag and the Bundesrat. Similarly, in the
United States, legislation needs the approval of the House and the Senate Chambers
and then it needs to be signed by the President. Political institutions can therefore not
single-handedly adopt legislation that corresponds to their preferred positions, but
they have to bargain with other institutional actors that have a say in the legislative
process. Legislative bodies that oppose a policy proposal can draw on three different
strategies. First, they can try to amend the proposal in their interest. Amending a
proposal however usually requires the consent of other institutional bodies which
might be hard to gain if they have opposing views. Second, political institutions
which have the right to veto legislation can block the adoption of the proposal by
not giving their approval. Finally, political actors can delay the adoption of a legis-
lative proposal. As the political composition of legislative chambers changes with
Klu
¨ver and Sagarzazu 391
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
regular elections, political actors which oppose a proposal can attempt to delay the
policy-making process until the next election. In addition, even if there is no election
coming up, delaying the decision-making process is a very promising strategy if time
is valuable to supporters of the proposal. In such situations, supporters of the legis-
lative initiative will make concessions in order to achieve timely agreement.
Duration of policy-making therefore crucially depends on the ideological con-
gruency of legislative bodies. If political institutions have similar preferences,
there is no need for extensive bargaining that delays the legislative process.
Legislative bodies largely agree and policy proposals are adopted quickly.
However, if legislative bodies have very distant ideal points, the degree of conflict
over a policy proposal is high. If policy preferences are very diverse, it is difficult to
find common ground and opponents attempt to block the adoption of a proposal.
Legislative bodies have to engage in extensive bargaining which usually involves
multiple package deals and concessions. Even if the consent of a legislative body
is not necessary, it can still delay the legislative process if its involvement is
mandatory.
To illustrate this point, we draw on the EP as an example. Even though the
consent of the EP is not required for the adoption of proposals subject to the
Consultation procedure, it can delay the legislative process by referring legislative
proposals back to the drafting committees (Kardasheva, 2009). The Consultation
procedure formally only consists of one reading, but the EP can gain an additional
reading by delaying a proposal. The EP can delay a proposal by putting it to vote
in the plenary in which MEPs can vote to refer the proposal back to the responsible
committee. The same proposal then has to be discussed again later in the plenary
when the EP issues its final opinion. What is more, the EP cannot only simply delay
a proposal, but it can also gain considerable influence on the design of the final
legislative act. If the EC and member states need an urgent decision on a proposal,
they have an incentive to speed up the legislative process by offering concessions to
the EP (Kardasheva, 2009).
The duration of policy-making thus crucially depends on the preference config-
uration of the political actors whose involvement in the legislative procedure is
mandatory. The larger the ideological diversity, the stronger the disagreement and
the longer it takes on average to adopt a legislative act. However, it is not reason-
able to simply count the number of all actors that are involved in the legislative
procedure and to calculate the distance between all of them in order to predict the
duration of policy-making. To explain the decision-making speed, it is crucial to
identify the legislative bodies that hold the most extreme policy positions. What
counts is only the maximum distance between the most extreme political actors.
Political actors that are located in between the most extreme legislative bodies do
not make a difference as they do not affect the set of possible outcomes. All pol-
itical institutions that have an ideal point which falls between the policy preferences
of the most extreme legislative bodies are ‘‘absorbed’’ by those bodies (Tsebelis,
2002: 26–29). We therefore argue that decision-making speed can be explained by
the maximum ideological distance between legislative bodies whose involvement is
392 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
mandatory for the adoption of a legislative act. The following hypothesis can be
formulated.
H1: The larger the ideological distance between the legislative bodies that hold
the most extreme policy positions, the longer the duration of the policy-making
process.
Research design
In this section, we illustrate how we constructed the dataset that was used to
empirically test our theoretical claim. We first discuss the measurement of deci-
sion-making duration. Afterwards, we illustrate how we operationalized the ideo-
logical distance between legislative bodies. Finally, we explain the measurement of
control variables.
Measuring duration
In order to measure decision-making speed, we draw on the EU database PreLex
which monitors the legislative process in the EU. With the aid of the multi-
threaded application LawLeecher, we downloaded all the legislative initiatives
included in this database until the 11th of November 2010 (Vogel et al., 2010).
4
As the first direct election to the EP only took place in 1979, we limit the analysis to
all legislative proposals introduced by the EC between 1979 and 2010. Since this
study aims at explaining variation in the duration of the legislative process, we only
concentrate on proposals for legislative acts that are subject to formal legislative
decision-making procedures (directives, regulations, decisions and recommenda-
tions). While regulations are directly applicable, directives have to be transposed
into national law. Decisions are directed at very specific recipients and recommen-
dations suggest a line of action without imposing any legal obligation. Table 1
provides an overview of the proposals in our sample which contains 12,093 legis-
lative acts. We therefore draw on the most comprehensive dataset to study duration
of legislative processes that has been presented so far.
To measure the duration of the decision-making process, we obtain the date
when the initiative was introduced and the date when the initiative was finally
adopted from the PreLex database. In order to operationalize the duration of
the legislative process for each policy initiative, we proceeded as follows. First,
we took the last date associated with a policy proposal and obtained the event
that took place at that particular date from all of the available dates associated
with a legislative proposal. If the date corresponds to the final adoption of the
proposal by the Council and if applicable the EP, we code the proposal as adopted
(1), otherwise we leave it as unapproved (0). If proposals are for instance adopted
under the Consultation procedure, the date that we used to generate decision-
making duration is the adoption by the Council (‘‘Formal adoption by
Council’’). If a proposal is by contrast subject to the Codecision procedure, the
date that we used to compute decision-making duration is the signature by the EP
Klu
¨ver and Sagarzazu 393
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
and the Council (‘‘Signature by EP and Council’’). From the 12,093 initiatives only
770 have not been adopted which shows that most of the initiatives proposed by the
EC eventually gain the consent of the Council and the EP. The speed with which
these initiatives have been adopted varies substantially. About 7 percent of initia-
tives are adopted in fewer than two weeks and more than 20 percent take more than
a year to be approved (see Table 2).
Measuring the ideological location of EU bodies
In order to operationalize the ideological congruency in the EU, we compiled data
on the ideal points of the EC, the Council and the EP as these three institutions are
decisive for the duration of the legislative process in the EU. There are three pri-
mary legislative procedures in the EU: Consultation, Cooperation and
Codecision.
5
The decision-making stage begins with the formal adoption of the
legislative proposal by the EC. The Council, the EP and the EC bargain about the
final legislative act on the basis of the Commission proposal. The Consultation
Table 2. Duration of legislative proposal approval
Time to approve Frequency Percent Cumulative percent
<2 weeks 824 7.28 7.28
2 weeks–1 month 1,245 11.00 18.27
1–2 months 1,668 14.73 33.00
2–3 months 1,203 10.62 43.63
3–4 months 1,076 9.50 53.13
4–8 months 1,982 17.50 70.63
8–12 months 999 8.82 79.46
1–2 years 1,389 12.27 91.72
2–5 years 754 6.66 98.38
>5 years 183 1.62 100.00
Total 11,323 100.00
Table 1. Legislative initiatives by type
Legislative type Frequency Percent
Decision 3,291 27.21
Directive 1,636 13.53
Recommendation 56 0.46
Regulation 7,110 58.79
Total 12,093 100.00
394 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
procedure requires that the EP is consulted, but the Council is not bound by the
Parliament’s opinion. In the Cooperation procedure, the Council can only overrule
the EP by unanimity. Finally, the Codecision procedure puts the Council and the
EP on equal footing as the Council cannot adopt a legislative act without the
approval of the Parliament.
We include the policy positions of all three legislative bodies for the following
reasons. The EC is responsible for the initiation of legislation and therefore enjoys
the monopoly of legislative initiative under the European Community Treaty
which regulates the vast share of EU policies. It is therefore crucial to take into
account the ideological position of the EC as its proposals set the terms for the
legislative debate between the Council and the EP. As the consent of the Council is
required for the approval of all policy proposals, it is also necessary to consider its
ideal point as no proposal can enter into force without its assent. Finally, even
though the consent of the EP is not required for all legislative proposals,
Kardasheva (2009) has shown that the EP can delay legislation even in situations
in which it does not have veto power by referring legislative proposals back to the
drafting committees. We therefore also take into account the ideological stance of
the EP.
Some observers argue that only the Council and the EP matter for legislative
outcomes under Codecision as the approval of the Commission is not required for
the adoption of a proposal after the conciliation committee has been convened.
However, the EC plays a crucial role for the decision-making speed as its proposals
forms the basis for the negotiations between the EP and the Council and since it is
formally involved in several instances throughout the process which have an effect
on decision-making duration (see also Burns, 2004). The Commission’s opinion on
the EP’s amendments in the second reading determines whether the Council can
adopt them by qualified majority or only by unanimity. What is more, the decision
of the EC whether to incorporate EP amendments into its proposal determines
whether the legislative process can be concluded after the first reading or whether a
Common Position has to be adopted which launches the second reading. Finally,
the EC also has an implicit effect on decision-making speed as the design of the
proposal affects whether amendments are introduced. If the Commission is largely
in line with the Council and the EP, the legislative proposal can be quickly adopted
in the first reading without any amendments. However, if the Commission presents
a proposal that is largely opposed by the EP and the Council, the legislative process
will take much longer as amendments will be suggested that delay the adoption of
the legislative act.
In order to measure the ideological distance between the European institutions,
we empirically assess their preferences on the left–right scale as several scholars
have concluded that the EU politics are characterized by a left–right space (Gabel
and Hix, 2002; Hix et al., 2005; Kreppel and Tsebelis, 1999; Tsebelis and Garrett,
2000). In addition, even though some other scholars disagree and suggest that the
European political space is characterized by a two-dimensional space consisting of
a left–right and a pro-anti European integration dimension (McElroy and Benoit,
Klu
¨ver and Sagarzazu 395
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
2007), there is hardly any variation in the pro-anti European integration dimension
as Warntjen (2008: 1248) have demonstrated.
6
We measure the ideological distance between the European institutions drawing
on policy position estimates that we computed based on election manifestos coded
by the Comparative Manifesto Project (CMP) (e.g. Budge et al., 2001; Klingemann
et al., 2006). Election manifestos are an ideal source of policy position estimates as
political parties clearly spell out their ideological stance on a variety of policy issues
(for measuring positions in the EU, see also Klu
¨ver and Rodon, 2012; Veen, 2011;
Warntjen et al., 2008). We employ the widely used ‘‘RILE’’ scale to estimate the
policy positions of legislative bodies in the left–right space. Human coders divided
the party manifestos into quasi-sentences and coded them according to a categor-
ization scheme consisting of left, right and neutral categories. The RILE scale is
computed as follows. First, the percentages of left and right categories of the total
number of coded quasi-sentences are computed. Then, the percentage of left sen-
tences is subtracted from the percentage of right sentences. Negative scores repre-
sent left positions and positive scores represent right positions. While the CMP
approach has been criticized for a lack of reliability, its superiority in terms of
validity over other computer-based text analysis approaches is widely acknowl-
edged. In addition, the availability of time-series position data constitutes a great
advantage over expert surveys.
Following Warntjen et al. (2008), we compute the ideal points of the three EU
bodies by using a weighted means approach. We estimate the ideal point of the EP
by taking the mean of the ideal points of all the national parties represented in the
EP as they are the driving forces behind voting behavior of MEPs weighted by their
number of seats. We generate the ideal point of the Council by taking the mean of
the policy positions of all governments represented in the Council weighted by their
number of votes. In case national governments were made up of coalitions, we
computed the government’s ideal point by weighting the policy positions of all
coalition parties by the share of cabinet seats that each coalition member holds.
Finally, the ideal point of the EC was estimated by taking the mean of the ideal
points of all Commissioners represented in the College of Commissioners. The
ideal points of Commissioners were based on the policy positions of the national
parties they are affiliated with. While this approach allows for estimating ideal
points for the EU institutions at any point in time, it implies an increase in meas-
urement error first induced by measuring national party ideal points by content
analysis and second intensified by aggregating various national party positions into
one single ideal point for the Commission, the Council and the EP. However, as it
is plausible to assume that the ideal points are not all systematically biased to the
left or to the right, the approach should deliver a good proxy of the true ideal
points of the European institutions. Figure 1 illustrates how the ideological com-
position of the European institutions changes over time.
To estimate the maximum distance between the European institutions we pro-
ceed in three steps. First, we identify all the unique ideological configurations of the
three legislative bodies in which none of the EU institutions undergoes a change in
396 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
its ideological composition. Unique ideological configurations refer to all instances
in time with stable ideal points of all three legislative bodies. Thus, neither the
ideological composition of the EC, nor the partisan composition of the Council
and the EP changes within these time intervals. We find that from 1979 to 2010
there are 71 such instances. Second, for each of those unique instances we identify
the institutions that are located at the two extremes of the ideological space. As a
third and final step, we compute the squared distance between the two legislative
bodies located at the extremes. Figure 2 shows the evolution of the distance
between the European institutions over time.
Control variables
Based on the findings of previous duration studies, we include several control
variables in the analysis (for reviews of the findings, see Golub, 2008; Ko
¨nig,
2008). First, we control for the extent of EP involvement as previous studies have
found that it slows down the decision-making process (Golub, 1999, 2002, 2007;
Golub and Steunenberg, 2007; Ko
¨nig, 2007; Schulz and Ko
¨nig, 2000). We extract
this information from the PreLex database based on the indication of the legisla-
tive procedure and from information provided about the actual involvement of the
EP during the legislative process. Following standard practice in the literature, we
operationalize formal EP involvement using a dummy variable distinguishing
between legislative proposals subject to Codecision and Cooperation and other
Figure 1. Ideological composition of the European institutions.
Klu
¨ver and Sagarzazu 397
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
proposals (see e.g. Ko
¨nig, 2007; Schulz and Ko
¨nig, 2000). Second, we control for
the type of legislative acts as Schulz and Ko
¨nig (2000) have shown that directives
take longer than regulations or decisions. Third, in order to take into account that
the timely adoption of proposals might be important to legislative bodies, we also
control for the number of days left in the respective terms of the three EU institutions
(Kovats, 2009). For the EC and the EP, we include the number of days left in their
legislative term. For the Council, we include the number of days until the end of the
Council Presidency which organizes and sets the agenda of the Council.
Fourth, we also control for the number of EU member states assuming that the
more member states are involved, the longer it should take to approve legislation.
As the Treaty of Amsterdam introduced the possibility of bringing Codecision
dossiers already to a conclusion at the end of the first reading, we included a
dummy variable that distinguishes between all proposals introduced before and
after the Treaty of Amsterdam (Toshkov and Rasmussen, 2012). Fifth, we include
legislative backlog in the empirical analysis to control for the legislative workload
of the EU institutions. We operationalize this variable by taking the number of
pending legislative proposals. Another variable that has attracted wide-spread
attention is the voting rule in the Council. Previous studies have found that
decision-making speed is faster when the Council decides by qualified majority
rather than by unanimity (Golub, 1999, 2002, 2007; Golub and Steunenberg,
2007; Ko
¨nig, 2007; Schulz and Ko
¨nig, 2000). However, for 10,650 out of 12,093
policy proposals, the PreLex database does not contain information about the
Figure 2. Maximum distance between the European institutions.
398 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
voting rule that applies in the Council. Accordingly, the dataset presented by Ha
¨ge
(2011) derived from the PreLex database suffers from a similar problem. We there-
fore refrain from including the voting rule in our base model. However, in order to
test the robustness of the findings, we estimate an alternative model specification
based solely on the subset of the data for which PreLex provides information about
the voting rule (see Web Appendix).
In addition, we control for two further institutional features. First, we control
for the asymmetry of powers between different legislative bodies. The Council, the
EP and the EC have different legislative powers and the effect of ideological diver-
sity might vary depending on which EU institution is absorbed. Finally, member
states play a crucial role in EU decision-making as shown in previous research (e.g.
Thomson, 2011). We therefore control for ideological diversity among member state
governments in the Council by including the standard deviation of the policy pos-
itions of the different member state governments weighted by their voting power in
the Council (see Ko
¨nig, 2007; Toshkov and Rasmussen, 2012).
Data analysis
After presenting our theoretical argument and illustrating the construction of the
dataset, we now turn to the empirical test of our theoretical claim. We first discuss
the specification of the statistical model that we use to empirically evaluate our
hypothesis. We then briefly discuss how we prepared the dataset to correspond to
the specific data structure of duration models before turning to the data analysis
and the presentation of the results.
Given the theoretical claims presented earlier and the nature of our data we will
rely on an event history model to analyze the effect of ideological congruency on
decision-making speed. Event history models are used in studies where the phe-
nomenon of interest is duration-to-event, such as duration until the adoption of a
policy proposal. Event history models allow us to take into account the specificities
of duration data that are typically characterized by censored observations. As we
do not have any a priori assumptions about the specific probability distribution for
the time until a proposal is adopted, we estimated a semi-parametric Cox model as
is standard practice in political science (see e.g. Box-Steffensmeier and Zorn, 2001;
Golub and Steunenberg, 2007). To take into account state changes in the dataset by
using time-varying covariates (TVCs), we first prepared our dataset accordingly.
TVCs are variables that undergo state changes so that the values initially assigned
to a case change over its lifetime (Golub, 2007: 163). State changes can mainly
occur with regard to our main explanatory variable as the ideological composition
of the European institutions can vary with the legislative term of the College of
Commissioners, the MEPs and the occurrence of national elections that bring new
governments into the Council. We therefore first identified all the changes in the
ideological composition of the three legislative bodies and then matched the
changes in the ideal points of the European institutions with our legislative pro-
posal database.
Klu
¨ver and Sagarzazu 399
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
An important assumption underlying the semi-parametric Cox model is that the
hazard function of each observation follows exactly the same pattern over time
(Box-Steffensmeier and Zorn, 2001: 974–975). Violations of this assumption might
result in biased coefficient estimates and decreased power of significance tests (Box-
Steffensmeier and Zorn, 2001: 974). We therefore tested this assumption on the
basis of Schoenfeld residuals (see Web Appendix). Eight of our independent vari-
ables violate the proportional-hazards assumption (EP involvement,days left in the
Commission term,days left in the Council Presidency term,directives,member states,
legislative backlog, the dummy for the Council as the absorbed body and intra-
Council diversity). Based on the results of the proportional-hazards test, we esti-
mate a semi-parametric Cox duration model with time-varying coefficients in order
to correct for the violation of the proportionality assumption as suggested by Box-
Steffensmeier and Zorn (2001). Including time-varying coefficients allows for taking
into account that the effects of covariates change over time. We use the natural
logarithm of time (ln(t)) to operationalize the non-proportional effect of the inde-
pendent variables. Table 3 presents the results of the model. The effects are
reported as raw coefficients and as hazard ratios. The hazard rate is an exponential
function of the coefficients and variables. As such, raw coefficients with a negative
value indicate a positive effect on duration while positive values indicate a negative
effect. Correspondingly, hazard ratios with a value between 0 and 1 indicate a
positive effect on duration while hazard ratios with a value above 1 indicate a
negative effect on duration.
The results of the Cox regression reported in Table 3 confirm our theoretical
expectation. The ideological distance between the most extreme legislative bodies
crucially affects the duration of legislative decision-making in the EU. The effect is
statistically significant and its direction is as expected. The larger the maximum
ideological distance between the three institutional bodies involved in policy-
making, the longer it takes to approve legislation. It is important to note that
while the small size of the coefficient might signal a small effect for this variable,
in reality this has to be seen in context. This coefficient signals that when the
two most extreme bodies are very close, the effect of this variable on increasing
the duration is negligible. However, as the distance between the most extreme
legislative bodies rises, the effect on increasing the time it takes to approve legis-
lation also increases. To further investigate the size of the effect of maximum
ideological distance we estimated the predicted marginals for our hazard function
when the distances were at their minimum and maximum values. We find that the
marginal hazard ratio, the rate at which bills get approved (die), is almost twice as
large when going from the minimum to the maximum distance (¼1.798,
s¼0.249 for the minimum distance and ¼1.056, s¼0.204 for the maximum
distance). In addition, we estimated the effect of a one standard deviation increase
in ideological distance from the mean distance to further illustrate the size of the
effect. As the ideological distance between the most extreme legislative bodies
increases by one standard deviation from its mean value, the hazard ratio rises
by one tenth. Thus, when the ideological distance increases by one standard
400 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
deviation from the mean, the speed at which bills are approved diminishes by
one tenth.
7
The effect of inter-institutional ideological distance on decision-making speed is
further illustrated in Figure 3. This figure shows the values for the cumulative
hazard function (Cleves et al., 2008) based on the smallest and largest values for
the maximum distance between the EC, the Council and the EP. The y-axis shows
the total number of legislative initiatives that have been approved and the x-axis
shows the number of days elapsed since the introduction of the initiatives. Figure 3
demonstrates that when maximum distance is at its smallest value (solid line), the
number of legislative initiatives that have been approved is about twice the number
of when the maximum distance (dashed line) is at its highest value.
With regard to the control variables, the results of the Cox regression reported
in Table 3 largely corroborate previous findings. In line with earlier research that
has found that the formal participation of the EP in the Cooperation and
Codecision procedure slows down the decision-making process (Golub, 1999,
Table 3. Cox regression explaining decision-making speed
Variable Coefficient Hazard ratio
Standard error
(coefficient)
Maximum distance 0.001*** 0.999*** 0.000
EP Involvement 5.381*** 0.005*** 0.206
Days left in EP term 0.000*** 1.000*** 0.000
Days left in Commission term 0.000*** 1.000*** 0.000
Days left in Council Presidency term 0.009*** 0.991*** 0.001
Directives 4.091*** 0.017*** 0.196
Member states 0.008 0.992 0.008
Amsterdam 0.208*** 1.231*** 0.049
Legislative backlog 0.001*** 1.001*** 0.000
Absorbed body – Commission 0.180*** 0.835*** 0.034
Absorbed body – Council 0.339*** 0.712*** 0.082
Intra-council diversity 0.073*** 1.076*** 0.009
EP Involvement ln(t) 0.852*** 2.345*** 0.035
Days left in Commission term ln(t)0.000*** 1.000*** 0.000
Days left in Council Presidency term ln(t) 0.001*** 1.001*** 0.000
Directive ln(t) 0.606*** 1.833*** 0.033
Member states ln(t) 0.003 1.003 0.002
Legislative backlog ln(t)0.000*** 1.000*** 0.000
Absorbed body – Council ln(t) 0.036** 1.037** 0.016
Intra-council diversity ln(t)0.010*** 0.990*** 0.002
N¼39,408, Subjects ¼12,081, Failures ¼11,192, Log likelihood ¼93,515.
***p0.01, **p0.05, *p0.10.
Klu
¨ver and Sagarzazu 401
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
2002, 2007; Golub and Steunenberg, 2007; Ko
¨nig, 2007; Schulz and Ko
¨nig, 2000),
we find that the formal involvement of the EP in the legislative process increases the
time it takes for a proposal to be approved. Similarly, our results also corroborate
the findings of Schulz and Ko
¨nig (2000) as it takes longer to approve directives
than other legislative proposals. The number of days left until the end of terms of
office of the three EU institutions indicate that at the beginning of the legislative
terms of the EC and the EP, legislative acts are approved more quickly than
towards the end of their terms. By contrast, when it comes to the Council, the
analysis shows that decision-making speed is faster towards the end of a Council
Presidency. The analysis furthermore shows that the opportunity of bringing
Codecision dossiers already to a conclusion at the end of the first reading which
was introduced in the Treaty of Amsterdam has lead to an acceleration of legisla-
tive decision-making. The Cox regression moreover indicates that legislative back-
log increases decision-making speed. In addition, the analysis shows that the
absorption of the EC and the Council as compared to the absorption of the EP
generally slows down policy-making. Finally, ideological diversity within the
Council has a positive effect on decision-making speed. This finding is surprising
as it signals that ideological diversity among member state governments leads cet-
eris paribus to faster legislative policy-making. Future research needs to further
investigate this puzzling finding.
In order to test the robustness of the findings, we estimated three further model
specifications (see Web Appendix). Due to data limitations, all three models are
Figure 3. Cumulative hazard function for the Cox regression of time for bill proposal
approval.
402 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
based on subsets of the full dataset. At first, we estimated a model with a more fine-
grained classification of EP involvement by distinguishing between Consultation,
Cooperation and Codecision. The results confirm our major claim: The ideological
distance between the two most extreme legislative bodies has a statistically signifi-
cant delaying effect on decision-making speed. In addition, the analysis corrobor-
ates previous findings as it is shown that decision-making takes longer if
Cooperation or Codecision apply. The second model checks the robustness of
the findings with regard to the ideological space. One could argue that prior to
the Treaty of Maastricht the debate centered on the degree of integration rather
than on left–right issues. We therefore estimated a second model on the basis of the
pro-anti European integration dimension for all legislative proposals that were
introduced before the Treaty of Maastricht entered into force. The results similarly
confirm our findings as the maximum distance between the EU institutions on the
pro-anti European integration dimension also systematically slows down decision-
making. Finally, as previous studies have shown that the voting rule has a major
impact on policy-making speed, we also estimated an additional model including
voting rule as a control variable. In line with previous findings, the analysis indi-
cates that decision-making under qualified majority voting is faster than decision-
making under unanimity in the Council.
8
Conclusion
How can decision-making speed be explained? We have shown in this study that
inter-institutional ideological congruence affects the duration of the decision-
making process in the EU. If political institutions have similar preferences, there
is no need for extensive bargaining that delays the legislative process. Legislative
bodies largely agree and policy proposals are quickly adopted. However, if legis-
lative bodies have very distant ideal points, the degree of conflict over a policy
proposal is high. In such situations, it is difficult to find common ground and
opponents attempt to block the adoption of the proposal. Legislative bodies
have to engage in extensive bargaining which usually involves multiple package
deals and concessions. We have furthermore demonstrated that legislative bodies
matter for the duration of policy-making processes even if they do not enjoy a
formal veto right. Thus, even though institutional actors might not be able to block
the adoption of legislation, they are able to slow down the decision-making process
if their involvement is mandatory. Slowing down the adoption of legislation is a
powerful tool in particular when time is valuable to decision-makers. Opponents of
a particular piece of legislation can thereby demand concessions by threatening to
delay the legislative process.
What matters is however not the ideological distance between all legislative
bodies involved in the decision-making process. If political institutions hold
policy positions that fall between the ideal points of more extreme legislative
bodies, they are absorbed as they do not affect the set of possible legislative
outcomes. The maximum ideological distance between the legislative bodies that
Klu
¨ver and Sagarzazu 403
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
hold the most extreme policy positions is therefore decisive for the duration of
legislative processes. While this study has made an important contribution in show-
ing that the maximum ideological distance plays a crucial rule for duration of legis-
lative decision-making, there are two issues that could not have been addressed here
due to methodological difficulties. First, we could not empirically assess the relation-
ship between ideological diversity and the location of the status quo and second, we
were not able to evaluate the impact of issue linkages on decision-making speed.
Future research therefore needs to further enhance our study by incorporating the
status quo as well as issue linkages into the empirical analysis of duration of policy-
making.
Our findings have furthermore important implications for the nature of political
conflict in the EU and the output legitimacy of political systems more generally. The
results of our analysis indicate that ideology plays a crucial role in European politics
as the ideological composition of the European institutions has a systematic effect on
decision-making speed. If the EC, the EP and the Council have opposing ideological
views, bargaining takes significantly longer than in situations where there is no dis-
agreement. Given that we have shown that the ideological distance between legisla-
tive bodies slows down the adoption of legislation, our findings also have important
implications for the output legitimacy of political systems. As modern policy-making
requires increasingly fast reactions, the problem-solving capacity of political systems
crucially depends on the speed with which legislation is adopted. Party competition
resolving around ideological orientations slows down the policy-making process as
opponents systematically extend the duration of legislative processes. Party compe-
tition therefore undermines the output legitimacy of a political system by causing
delays in the adoption of legislation.
Notes
1. The authors follow alphabetical order. Both authors have contributed equally to the
article.
2. Toshkov and Rasmussen (2012) additionally limit the analysis of inter-institutional dis-
tance to the distance between the EP and the Council while ignoring the Commission and
also not taking into account that only the maximum distance between the three EU
institutions matters as it is argued in this study.
3. While we treat legislative bodies as unitary actors, we acknowledge the crucial role of
member states in the EU and therefore take into account heterogeneity within the
Council as a control variable in the empirical analysis.
4. The LawLeecher is available at: http://lawleecher.sourceforge.net/
5. The Treaty of Lisbon renamed the ‘‘Codecision’’ procedure the ‘‘Ordinary legislative
procedure’’. However, since the vast majority of proposals analyzed in this study was
approved before the Treaty of Lisbon entered into force, we use the term ‘‘Codecision’’
throughout the article.
6. However, as much of the legislation discussed in the 1980s and early 1990s dealt with the
transfer of competences from the national to the European level, we additionally tested
our theoretical claim relying on the pro-anti European integration dimension for this time
period (see Web Appendix).
404 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
7. Holding all other independent variables constant at their mean or median values (for
dichotomous and categorical variables).
8. We have performed further robustness checks by dropping different control variables
from the analysis. In addition, given that the ideological distance between the EU insti-
tutions was most pronounced between 1980 and 1985, we also ran a model excluding all
proposals introduced in this period. Across all these model specifications, the effect of the
ideological distance remains stable.
Acknowledgements
We thank Ernesto Calvo, Fabio Franchino, Christine Reh, Toni Rodon, Andreas Warntjen,
Gerald Schneider and the anonymous reviewers for valuable comments and suggestions.
Funding
This research received no specific grant from any funding agency in the public, commercial,
or not-for-profit sectors.
References
Box-Steffensmeier JM and Zorn CJW (2001) Duration models and proportional hazards in
political science. Americal Journal of Political Science 45(4): 972–988.
Budge I, Klingemann HD, Volkens A, Bara J and Tanenbaum E (2001) Mapping Policy
Preferences: Estimates for Parties, Electors, and Governments 1945–1998. Oxford: Oxford
University Press.
Burns C (2004) Codecision and the European commission: A study of declining influence?
Journal of European Public Policy 11(1): 1–18.
Calvo E and Sagarzazu I (2011) Legislator success in committee: Gatekeeping authority and
the loss of majority control. American Journal of Political Science 55(1): 1–15.
Cleves M, Gould W, Gutierrez R and Machenko Y (2008) An Introduction to Survival
Analysis Using Stata. College Station, TX, USA: Stata Press.
Cox GW and McCubbins MD (1993) Legislative Leviathan: Party Government in the House.
Berkeley, CA, USA: University of California Press.
Downs A (1957) An Economic Theory of Democracy. New York: Harper.
Do
¨ring H (1995) Time as a scarce resource: Government control of the agenda. In: Do
¨ring H
(ed.) Parliaments and Majority Rule in Western Europe. Frankfurt: Campus, pp. 223–246.
Dru
¨ner D, Klu
¨ver H, Mastenbroek E and Schneider G (2012) The Core or the Winset?
Explaining Decision-Making Duration and Policy Change in the European Union.
Working Paper.
Gabel M and Hix S (2002) Defining the EU political space: An empirical study of
the European elections manifestos, 1979–1999. Comparative Political Studies 35(8):
934–964.
Goetz KH (2009) How does the EU tick? Five propositions on political time. Journal of
European Public Policy 16(2): 202–220.
Golub J (1999) In the shadow of the vote? Decision making in the European Community.
International Organization 53(4): 733–764.
Golub J (2002) Institutional reform and decisionmaking in the European Union. In: Hosli M
and van Deemen A (eds) Institutional challenges in the European Union. London, UK:
Routledge, 134–154.
Klu
¨ver and Sagarzazu 405
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
Golub J (2007) Survival analysis and European Union decision-making. European Union
Politics 8(2): 155–179.
Golub J (2008) The study of decision-making speed in the European union: Methods, data
and theory. European Union Politics 9(1): 167–179.
Golub J and Steunenberg B (2007) How time affects EU decision-making. European Union
Politics 8(4): 555–566.
Hagemann S and Høyland B (2008) Parties in the Council? Journal of European Public
Policy 15(8): 1205–1221.
Hix S (2002) Parliamentary behavior with two principals: Preferences, parties, and voting in
the European Parliament. American Journal of Political Science 46(3): 688–698.
Hix S (2004) Electoral institutions and legislative behavior: Explaining voting defection in
the European Parliament. World Politics 56(2): 194–223.
Hix S, Noury A and Roland G (2005) Power to the parties: Cohesion and competition in the
European Parliament, 1979–2001. British Journal of Political Science 35(2): 209–234.
Howell W, Scott Adler ACC and Riemann C (2000) Divided government and legislative
productivity of congress, 1945–1994. Legislative Studies Quarterly 25: 285–312.
Ha
¨ge F (2011) The European Union policy-making dataset. European Union Politics 12(3):
455–477.
Kardasheva R (2009) The power to delay: The European Parliament’s influence in the
consultation procedure. Journal of Common Market Studies 47(2): 385–409.
Klingemann HD, Volkens A, Bara J, Budge I and McDonald M (2006) Mapping Policy
Preferences II: Estimates for Parties, Electors, and Governments in Eastern Europe,
European Union and OECD 1990–2003. Oxford: Oxford University Press.
Klu
¨ver H and Rodon T (2012) Explaining policy position choice of Europarties: The effect
of legislative resources. British Journal of Political Science. Epub ahead of print doi:
10.1017/S0007123412000543
Ko
¨nig T (2007) Divergence or convergence? From ever-growing to ever-slowing European
legislative decision making. European Journal of Political Research 46(3): 417–444.
Ko
¨nig T (2008) Analysing the process of EU legislative decision-making: To make a long
story short....European Union Politics 9(1): 145–165.
Kovats L (2009) Do elections set the pace? A quantitative assessment of the timing of
European legislation. Journal of European Public Policy 16(2): 239–255.
Kreppel A and Tsebelis G (1999) Coalition formation in the European parliament.
Comparative Political Studies 32(8): 933–966.
Mattila M and Lane JE (2001) Why unanimity in the Council? European Union Politics 2(1):
31–52.
McElroy G and Benoit K (2007) Party groups and policy positions in the European
Parliament. Party Politics 13(1): 5–28.
McElroy G and Benoit K (2010) Party policy and group affiliation in the European
Parliament. British Journal of Political Science 40(02): 377–398.
Moravcsik A (1998) The Choice for Europe: Social Purpose and State Power From Rome to
Maastricht. Ithaca: Cornell University Press.
Rasmussen A and Toshkov D (2010) The inter-institutional division of power and time
allocation in the European parliament. West European Politics 34(1): 71–96.
Sandholtz W and Stone Sweet A (1998) European Integration and Supranational Governance.
Oxford: Oxford University Press.
Schulz H and Ko
¨nig T (2000) Institutional reform and decision-making efficiency in the
European Union. American Journal of Political Science 44(4): 653–666.
406 European Union Politics 14(3)
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
Thomson R (2011) Resolving Controversy in the European Union: Legislative Decision-
Making Before and After Enlargement. Cambridge: Cambridge University Press.
Toshkov D and Rasmussen A (2012) Time to decide: The effect of early agreements on
legislative duration in the EU. European Integration Online Papers (EIoP) 16(article 11).
Available at: http://eiop.or.at/eiop/texte/2012-011a.htm
Tsebelis G (1995) Decision making in political systems: Veto players in presidentialism,
parliamentarism, multicamerialism and multipartyism. British Journal of Political
Science 25(3): 289–325.
Tsebelis G (1999) Veto players and law production in parliamentary democracies: An empir-
ical analysis. American Political Science Review 93(3): 591–608.
Tsebelis G (2002) Veto Players: How Political Institutions Work. New York: Sage.
Tsebelis G and Garrett G (2000) Legislative politics in the European Union. European Union
Politics 1(1): 9–36.
Tsebelis G and Money J (1997) Bicamerlism. Cambridge: Cambridge: University Press.
Veen T (2011) Positions and salience in European union politics: A new dataset. European
Union Politics 12(2): 267–288.
Vogel T, Kovats L and Go
¨tz KH (2010) LawLeecher: A software for data extraction from
PreLex. Potsdam: University of Potsdam.
Warntjen A, Hix S and Crombez C (2008) The party political make-up of EU legislative
bodies. Journal of European Public Policy 15(8): 1243–1253.
Klu
¨ver and Sagarzazu 407
at Glasgow University Library on May 19, 2015eup.sagepub.comDownloaded from
