https://doi.org/10.1080/07036337.2017.1332057
Do votes matter? Voting weights and the success probability
of member state requests in the Council of the European
Union
Andreas Warntjen
Department of public administration, university of twente, Enschede, the netherlands
ABSTRACT
The impact of votes on Council decision-making is a widely studied subject. A large body of theory supports the contention that more votes, ceteris paribus, translate into more influence. Empirical studies have, however, shown a negative relationship between the voting power of a member state and its bargaining success. Based on a new data set, this study shows empirically that there is a robust positive relationship between the number of votes backing a member state request to change European legislation and its success probability. The analysis is based on all unique member state requests for exemptions, lower standards or longer transitional time periods in the field of environmental policy between the first wave of Eastern Enlargement in 2004 and the entry into force of the Lisbon treaty in 2009 (N = 216). It also shows that member state requests are often successful, contrary to the image of an unresponsive Brussels.
Introduction
Legislative decisions made in Brussels have a large effect on European Union (EU) member states. Preventing legislation that adopts higher European standards or establishes a higher level of harmonisation than they envisage is a frequent concern of member state govern-ments. They might oppose Commission proposals because of ideological views, economic interests or electoral consideration. Member state governments tend to have less integra-tionist positions and favor less regulation than the European Commission, which drafts
European legislation (Thomson 2011). All legislation has to pass through the Council, giving
member states a chance to request amendments in line with their political preferences or economic interests. Formal voting in the Council might be rare. Nevertheless, legislation cannot be passed without the tacit approval of a sufficient majority in the Council
(Hayes-Renshaw and Wallace 2011). A group of countries that requests lower regulatory
standards and has sufficient votes to block legislation can credibly threaten to prevent any new legislation in that area. Thus, member states in favor of more regulation as well as the
© 2017 the author(s). published by Informa uK limited, trading as taylor & francis group.
this is an open access article distributed under the terms of the Creative Commons attribution-nonCommercial-noDerivatives license (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.
KEYWORDS Council of the European union; voting power; blocking minority; legislative decision-making; member state interventions; gridlock
CONTACT andreas Warntjen a.K.Warntjen@utwente.nl
the supplemental data for this article is available online at https://doi.org/10.1080/07036337.2017.1332057.
OPEN ACCESS
other legislative bodies would have to accept changes (e.g. lower standards, a limited scope, or longer transitional periods) to accommodate a potential blocking minority. That more votes, in general, translate into more influence is a widely shared proposition, particularly
in the literature based on rational choice theory (e.g. Felsenthal and Machover 1988; Mueller
2003). Thus, the success probability of member state requests should increase with the
number of votes backing a particular request. In particular, requests backed by a blocking minority should be more successful than requests supported by fewer votes In other words, votes should matter in Council negotiations on European legislation. However, several empir-ical studies found a negative relationship between voting power and bargaining success.
Using data on all legislative proposals on environmental policy discussed in the Council after Eastern enlargement in 2004 and before the coming into force of the Lisbon treaty in 2009, this study shows that member states routinely make requests that would lower the level of regulatory standards or European harmonisation. These requests are quite often successful within the Council but also affect the final outcome of the legislative process. A multilevel regression analysis shows a positive relationship between the number of votes backing a proposal and more specifically whether or not the supporters of a request con-stitute a blocking minority and its success probability. As an observational study, the analysis cannot definitively establish a causal connection or rule out the impact of alternative mech-anisms. Nevertheless, the results suggest that legislative proposals are adapted in the dis-cussions in the Council to secure a sufficient majority. Thus, within Council negotiations, if not beyond, votes do matter.
This article provides novel empirical insights based on an original data-set to several ongoing debates on EU legislative politics. Previous studies of member state interventions found no connection or even a negative relationship between the voting power of individual member states and bargaining success. In addition, they found that member states that put forward requests were less successful in the negotiations in the Council than less active
member states (Cross 2013; Arregui 2016). In contrast, this study shows that member state
requests that are supported by more votes are more likely to be successful. These studies are not directly comparable, because they use different variables, a different unit of analysis and different research designs. In particular, studies on bargaining success are based on a sample consisting only of salient and controversial proposals. The discrepancy of the empir-ical findings regarding the effect of voting weights and the success rate of member state requests in the Council is striking nevertheless. This is also relevant for the debate on the
Council’s decision-making mode (Lewis 2005; Naurin 2010; Warntjen 2010). According to
constructivist approaches, not votes but arguments matter in Council decision-making. The high success rate of member state requests is also quite interesting for the debate on a democratic deficit due to a missing connection between the concerns voiced at the national
level and decisions taken in Brussels (Hix 2008; Moravcsik 2008). In terms of the relationship
between intra-institutional and inter-institutional negotiations (Naurin and Rasmussen
2011), it is also intriguing that changes made during Council negotiations usually make it
into the final piece of legislation.
The impact of voting weights on legislative decision-making in the Council In the literature on decision-making in the Council, voting weights plays a prominent role
(Hayes-Renshaw and Wallace 2011; Hix and Hoyland 2011). Several studies have addressed
how voting power affects not just the overall likelihood of change but also the power of
individual member states (Felsenthal and Machover 1997, 2001; Koczy 2012). These
consid-erations also played a prominent role in negotiations on Council reform (Moberg 2002;
Norman 2005; König, Warntjen, and Burkhart 2006).
Voting power is based on both the distribution of voting weights and the adoption
thresh-old (Felsenthal and Machover 1988); thus, having more voting weights does not always lead
to more voting power. However, in general more votes increase the power of an actor. Bargaining models of legislative decision- making often incorporate (weighted forms of)
voting power (Bueno de Mesquita and Stokman 1994; Thomson et al. 2006; Schneider, Finke,
and Bailer 2010). Controlling more votes should also increase influence in models of vote
trading (Stratmann 1997; Aksoy 2012) because votes are the currency of log-rolling. In all of
these rational choice models, having more voting weights, ceteris paribus, translates into more influence over decision-making outcomes. In contrast, constructivism highlights other factors such as socialisation and deliberation. Arguments and norms rather than voting weights determine legislative decision-making according to constructivist approaches (Risse
2000; Lewis 2005; Naurin 2010; Warntjen 2010).
A number of empirical studies seem to contradict the rational choice proposition that
voting weights are positively related to influence. Mattila (2004) argues, in line with the
voting power literature, that states with more votes should be more influential and thus less likely to vote against a proposal. Analyses of pre-enlargement roll call data, however, showed that larger states tend to vote more often against a proposal, contrary to theoretical
expec-tations (Mattila and Lane 2001; Mattila 2004). In contrast, a more recent study shows the
expected negative relationship between voting weights and the propensity to vote against a proposal if one controls for the position of an actor. This finding, however, does not hold
for the new member states (Hosli, Mattila, and Uriot 2011). Several studies have addressed
the bargaining success of member states based on expert surveys (DEU) which provided information on the ideal position of a member state, the importance it attaches to an issue
and the final outcome (Thomson 2011). Bargaining success (or satisfaction) is defined in
these studies as the (salience-weighted) distance between the ideal position of an actor and
the final outcome of legislative deliberations (e.g. Arregui and Thomson 2009, 664). Using a
salience-weighted measure of bargaining success, studies either find no systematic
differ-ence across small and large member states (Thomson 2011) or report that small member
states are more successful (closer to the final outcome) than larger ones despite having less
votes (Arregui and Thomson 2009; Golub 2012; Cross 2013). Cross (2013) studied the effect
of member states intervening at the working group or COREPER level. He shows that inter-ventions either have no effect on the distance between the ideal position of an actor and the final outcome or make it less likely that an actor is satisfied with the final outcome if one considers the importance an actor attaches to an issue. His multivariate regression analysis controls for the relative position of member states vis-à-vis other member states, the status
quo and the positions of the other legislative actors (Cross 2013, Table 1). Using a data-set
which includes the post-enlargement period, Arregui (2016) also finds that voting power is
not related with bargaining satisfaction and that higher involvement (e.g. making requests for changes) has a negative effect. His multivariate regression analysis controls for the relative position of member states vis-à-vis the status quo, the European Parliament, the Commission as well as the extremity of their position within the Council. Thus, the existing empirical literature suggests that member states that have more votes and intervene more often are
not more successful than less powerful or active states. Even when controlling for the relative position of member states (vis-à-vis the Commission and European Parliament), the extremity of a member state position and the salience a member state attaches to an issue, voting power (or population size) has either no or a negative effect on bargaining success (Arregui
2016, Table 2; Arregui and Thomson 2009, Table 2; Cross 2013; cf. Golub 2012, 1308). These
surprising empirical findings have been reconciled in different ways with the theoretical
expectation of a positive effect of voting power. Mattila (2004) argues that the higher
fre-quency of larger member states voting against legislation might be explained by their expec-tation to be disproportionately influential. Larger member states might be more prone to voice their opposition when outvoted, whereas smaller member states accept their limited influence without formal protest. Similarly, member states might merely intervene to express
their discontent rather than to change negotiation outcomes (Cross 2013). Large member
states might also lose more often on issues, because they simply more often have a stake in
negotiations due to their more diversified economies (Arregui and Thomson 2009). In
addi-tion, smaller member states might exert more influence on those issues where they do have an interest because they can credibly claim that those issues really are vital for them (Arregui
2016). Finally, smaller member states might be more effective negotiators because they can
focus on fewer issues and find it easier to coordinate their position internally (Panke 2010).
Nevertheless, the lack of robust evidence for a positive relationship between the number of votes and bargaining success is puzzling in light of rational choice theories of EU deci-sion-making. The analysis of roll call votes, however, does not necessarily tell us how influ-ential a government actually was. Existing studies on bargaining success based on expert surveys only consider the final outcome of the tri-partite legislative negotiations. This con-flates the impact of votes in Council decision-making, where it should be strongest, with effects related to the negotiations of the Council with the supranational bodies. Previous studies are also based on a non-representative sample of legislative proposals. We can address the issue of the impact of votes in Council decision-making more directly by looking at the relationship between the number of votes and the success probability of member state proposals. This will also give additional empirical insights into the impact of member states interventions on negotiation outcomes.
The theoretical impact of votes in the Council can be illustrated with a spatial model
(Tsebelis 2002). Figure 1 shows the ideal points of four countries (A, B, C, D) in a two-
dimensional policy space. The horizontal axis denotes the level of regulation and the vertical axis represents the scope of a regulation. A is against any change of the status quo (sq) towards a higher level of regulation or a wider scope of EU legislation. The other three actors (B, C, D) differ with regard to the degree to which more regulation or a wider scope would be acceptable.
Let us assume that 75 (out of 100) votes are necessary to adopt new legislation and that all countries have the same number of votes, namely 25. Commission proposals tend to set
out a high level of regulation and harmonisation (Thomson 2011, 79–101). Consider the
situation where the Commission proposal is located at x. This proposal would fail to garner a sufficient majority because only two countries (C, D) with 50 votes would prefer it to the status quo. Changing the content of the legislation by restricting its scope, however, would yield a sufficient majority. For example, shifting the proposal to policy z would be acceptable to a coalition of three countries (B, C, D) which controls 75 votes. The effect of delaying the entry into force of a proposal by extending transitional time periods has a similar effect to
restricting the scope of a directive in terms of its applicability. In addition, adapting legislative proposals by lowering standards can be used to overcome possible gridlock. Consider the situation where the Commission has proposed policy y, which is only supported by two member states (B, D) with 50 votes. In contrast, a proposal with a lower level of regulation (e.g. policy z) would be acceptable to a sufficient majority (B, C, D). Member states who favor more far-reaching regulations might object to changes that lower standards, increase tran-sitional time periods or limit the scope of a European regulation. In a regulatory domain like environmental policy, they, however, would prefer a Europe-wide regulation at a lower level
to having no regulation at all (Scharpf 1996). For example, actor D in Figure 1 prefers policy
z to the status quo. In contrast, member state governments that prefer less regulation or less harmonisation can credibly threaten to vote against a proposal unless it is adapted in line with their request for a limited scope, lower standards or a longer transitional period. Thus, the success rate of member state requests should increase with the number of votes backing a particular request. In particular, requests backed by a blocking minority should be successful. A blocking minority consists of a coalition of actors who control enough votes to prevent a passage of a proposal. In our example, any coalition of two actors with 50 votes constitutes a blocking minority. For example, a blocking minority consisting of A and B would block proposal x. Similarly, a coalition consisting of A and C would block proposal y. In order to ensure passage of a proposal, the content of a proposal would have to be changed in line with a request for derogations, a delayed entry into force or lower standards (e.g. from x or y to z). Thus, requests backed by a sufficient number of votes to constitute a blocking minority should in general be more likely to be enacted than requests with less support.
Figure 1. Blocking minorities and the success probability of member state requests.
H1: Proposals backed by a blocking minority of member states have a higher probability of success than proposals backed by fewer votes.
Member state governments do not possess perfect information regarding the voting intentions of other member states. Not all governments backing a request will be willing to veto a piece of legislation over it, so this is only a probabilistic relationship with no sharp threshold right at the number of votes necessary for a blocking minority. The higher the number of member states backing a request for changes, however, the more likely it is that the whole proposal might be blocked. Thus, the probability of success of a request should increase with the number of votes backing a request.
H2: The probability of success of member states’ requests increases with the number of votes backing a request.
Data collection and operationalisation
In order to study the impact of votes on legislative decision-making in the Council, this study looks at the success probability of member state requests between the first round of Eastern enlargement (May 2004) and the entry into force of the Lisbon treaty (December 2009) in
one regulatory policy field (environmental policy). The data-set1 distinguishes between the
different phases of the legislative process and between partial and full success. The DEU-based studies cited above are restricted to salient and controversial proposals (Thomson
2011, 28). By looking at all member state requests in one policy field, this study can give us
a more representative picture of the effect of voting weights.
The EUPOL data-set (Häge 2011) was used to identify all legislative proposals (regulations
or directives) on which the Council had not reached political agreement in first reading before 1 May 2004 and which could have been discussed in the Council before the entry into force of the Lisbon treaty on 1 December 2009. The voting rules of the Nice Treaty applied throughout this period. Decisions were excluded because they only apply to a small set of subjects and thus would not necessarily be of interest to all member states. The data was gathered manually based on summaries of Council discussions (mainly at the working group level) and annotated proposals during first reading. The restriction to first reading is due to the focus on the Council. In some instances, the publicly available information is restricted
due to political controversy (Cross 2012). The annotated proposals are sometimes reflecting
changes not relative to the Commission proposal but to a presidency proposal. Thus, the requests in the data-set are effectively a subset of all member state requests for changes to the Commission proposal, some of which are already included in the proposal as amended by a presidency compromise proposal. Requests were identified for 23 out of 43 legislative proposals discussed in the Council.
The data-set consists of all requests by member states that were specific and not just related to procedural issues. The latter restriction excludes discussion on whether comitology would apply, in which format or frequency results would be reported to the Commission and public participation. Member state requests were coded into three separate categories: derogations, extensions and lower standards. Derogations are all requests that would reduce the scope of legislation, for example by making specific exemptions or by limiting the scope of activities to which the legislation applies. For example, several member states requested derogations to exclude flights by small planes, for public services and to remote areas from
the scope of the directive on the inclusion of aviation activities in the greenhouse gas
emis-sion trading scheme (European Commisemis-sion 2006). Extensions refer to requests that would
extend the time period before the legislation is effective (e.g. requesting a later date for transposition of directives or entry into force, requests for transitional time periods). Lower standards refers to requests that would lower standards of the proposal (e.g. by increasing limit values for pollution or increasing threshold values which would trigger mandatory actions to limit pollution). For example, several member states requested a higher limit value for some pollutants and longer transitional time periods in the Council discussions on the
directive on the quality of ambient air (European Commission 2005).
For each unique request the degree to which member states’ requests were successful was coded manually based on a comparison of the requests, the common position of the Council and the final adopted act. We can distinguish between partial and full success and success within the Council and across the whole procedure. A request was fully successful if the outcome completely incorporates the requested change. A request was partially suc-cessful if there were changes that reflect the request at least to some degree (e.g. an exten-sion of a transitional time period is included, even if not for the entire amount requested). The definition of partial success includes cases where requests are completely successful. Final success refers to the outcome of the complete procedure. If the proposal was adopted in first reading, we cannot distinguish between a position of the Council and the final out-come because there is no separate common position of the Council. This is the case for about one quarter of the unique requests. Thus we can distinguish between success regarding the Common Position of the Council, the entire first reading (including both Common Positions and procedures that were concluded in first reading), and the final outcome (which also includes procedures concluded after the first reading).
Data analysis
Member states frequently put forward requests to limit the scope of European legislation, extend
transitional time periods or lower standards (cf. Warntjen 2017). These requests are often
suc-cessful. The success probability increases with the number of votes backing a particular request. Member states routinely request changes to the proposals put forward by the Commission. Member states made 790 individual requests or 216 unique requests (identical individual requests made by several member states) regarding 43 legislative proposals. Requests were made for more than half of the proposals (23). Most of the proposals are put forward by several member states, which makes it difficult or impossible to estimate the effect of char-acteristics of individual member state governments (e.g. voting weights or the relative posi-tion of a government). The following analysis is based at the level of unique requests for these proposals, thus usually comparing the success of coalitions of member states rather
than individual member states. Table 1 shows the distribution of unique requests across
type of request and the type of legislation. The most common type of requests are for low-ering standards, closely followed by derogations. Requests for extensions are the least com-mon. Nearly a quarter of requests were backed by a blocking minority. The high voting
thresholds of the Nice treaty (Hosli and Machover 2004; Tsebelis 2008) were still in place in
the time period studied here. A blocking minority requires 90 (EU-25) or 91 (EU-27) votes. There is only one request supported by exactly 90 votes which falls into the EU-27 period. Hence, 91 votes is used as the threshold for a blocking minority.
The success rates of requests are similar across the different stages of the legislative procedure, but differ with regard to the type of request. The values of the success variables for the Council, first reading and the final outcome are highly associated. In 98% of these cases, the values of the partial success variable for the Council and the final outcome are identical. Similarly, in 97% of the cases where the Council adopted the request of a member
state fully in its Common Position this is still included in the final outcome. Table 2 shows
the success rates across the different stages of the legislative procedures and by type of request. Overall, 45% of requests are at least partially successful and about 30% are fully successful. Requests for extensions are considerably more often successful than requests for derogations or lower standards when considering partial success. The difference between type of requests is a lot less pronounced for complete success.
Proposals that are backed by a blocking minority are more successful than proposals which are supported by fewer votes (H1). Overall, 45% of requests were at least partially successful. Proposals that were put forward by a blocking minority were successful in 76% of the cases. In contrast, only 37% of the proposals without a blocking minority have been
successful. Table 3 compares the success rates for the different stages of the decision-making
Table 1. Descriptive statistics.
note: percentages are rounded values.
Frequency Percent Type of legislation Directive 181 83.8 regulation 35 16.2 total 216 100 New/amending amending 80 37 new 136 63 total 216 100 Type of request Derogation 78 36.1 Extension 54 25 lower 84 38.9 total 216 100 Blocking minority Yes 51 23.6 no 165 76.4 total 216 100
Table 2. Success rate (percent) across stages.
note: rounded values.
Derogation Extension Lower standards All
At least partial success
Common position 37.3 57.1 37.3 42.5 first reading 40.9 58.5 38.5 44.6 final act 42.7 56.9 36.4 43.9 all 40.9 58.5 38.5 44.6 Full success Common position 23.5 28.6 28.4 26.8 first reading 31 32.1 28.2 30.2 final act 31 27.5 28.6 29.1 all 31 32.1 28.2 30.2
procedure and the different degrees of success (at least partial, full). The number of cases differs because of missing values for the success variable. In some instances, the changes in a proposal were too complex to clearly identify if they were linked to a request. In addition, some proposals did not have a Common Position or the final decision was made after the entry into force of the Lisbon treaty. Proposals backed by a blocking minority are always
more successful. A χ2-test for independence confirms this with p < 0.01 for all settings except
full success for the final act.
A multivariate analysis confirms the relationship between blocking minorities and the
success probability of a request (Table 4). The data is grouped by proposal as several requests
are pertinent to the same proposal. A multilevel model was estimated which controls for the type of proposal (directive/regulation, amending/non-amending) and the number of recitals of the Commission proposal. The number of recitals is related to the salience and the complexity of a proposal and has also been used as a measure for controversy (Warntjen
2012). Proposals that are more controversial, important, and/or complex might attract more
attention and thus might be more likely to be adapted in general, hence it is useful to control for these factors in the multivariate analysis. Similarly, regulations might be subject to more Table 3. Blocking minorities and success rate.
note: row percentages are rounded values. p-values are for a χ2 test of independence. Success (percent)
Total
No Yes
(a) Partial, common position (N = 160)
no blocking minority 67.2 32.8 100
Blocking minority 26.3 73.7 100
total 57.5 42.5 100
p < 0.001
(b) Full, common position (N = 160)
no blocking minority 79.5 20.5 100
Blocking minority 52.6 47.4 100
total 73.1 26.9 100
p = 0.001
(c) Partial, first reading (N = 202)
no blocking minority 64.7 35.3 100
Blocking minority 26.5 73.5 100
total 55.5 44.6 100
p < 0.001
(d) Full, first reading (N = 202)
no blocking minority 75.8 24.2 100
Blocking minority 51 49 100
total 69.8 30.2 100
p = 0.001
(e) Partial, final (N = 196)
no blocking minority 64.6 35.4 100 Blocking minority 30.6 69.4 100 total 56.1 43.9 100 p < 0.001 (f) Full, final (N = 196) no blocking minority 75.5 24.5 100 Blocking minority 57.1 42.9 100 total 70.9 29.1 100 p = 0.014
Table 4.
the number of v
ot
es
, block
ing minorities and the pr
obabilit y of suc cess (multilev el log istic r eg ression). n ot e: S tandar d err ors in br ack ets . *p < 0.10; ** p < 0.05; *** p < 0.01. Common position First r eading Final ac t Par tial Full Par tial Full Par tial Full Block ing minorit y 1.918*** (0.478) 1.379*** (0.448) 1.749*** (0.409) 1.184*** (0.378) 1.558*** (0.398) 0.907** (0.392) n umber of vot es 0.024*** (0.005) 0.017*** (0.004) 0.021*** (0.004) 0.017*** (0.004) 0.018*** (0.004) 0.015*** (0.004) Ex tension 0.856* (0.510) 0.721 (0.523) 0.122 (0.538) 0.022 (0.549) 0.834* (0.445) 0.666 (0.454) 0.049 (0.451) −0.148 (0.467) 0.607 (0.449) 0.433 (0.455) −0.188 (0.489) −0.400 (0.505) lo w er standar ds −0.355 (0.487) −0.596 (0.507) 0.046 (0.502) −0.140 (0.518) −0.376 (0.413) −0.611 (0.432) −0.272 (0.418) −0.526 (0.439) −0.569 (0.417) −0.772* (0.432) −0.196 (0.433) −0.422 (0.451) amending −0.472 (0.856) −0.343 (0.829) −0.433 (0.764) −0.264 (0.736) −0.179 (0.507) −0.177 (0.496) 0.093 (0.469) 0.172 (0.460) −0.294 (0.495) −0.276 (0.481) 0.085 (0.531) 0.170 (0.514) Dir ec tiv e −0.666 (0.805) −0.845 (0.761) −1.073 (0.686) −1.316** (0.660) −0.447 (0.579) −0.507 (0.559) −0.558 (0.537) −0.737 (0.526) −0.836 (0.612) −0.839 (0.588) −1.111* (0.628) −1.248** (0.608) recitals 0.035 (0.043) 0.045 (0.042) −0.018 (0.036) −0.017 (0.035) 0.030 (0.022) 0.027 (0.022) 0.014 (0.020) 0.009 (0.020) 0.025 (0.022) 0.023 (0.021) 0.008 (0.023) 0.003 (0.022) In ter cept −1.189 −1.324 −2.071 −1.308 −0.116 −1.125 −0.610 −1.101 −1.224 (0.906) −1.787** (0.902) −1.135 (0.844) −1.499* (0.840) −0.573 (0.946) −1.095 (0.937) −0.513 (0.972) −0.867 (0.956) random eff ec t (in ter cept) −0.175 (0.346) −0.301 (0.400) −0.523 (0.495) −0.698 (0.623) −0.347 (0.340) −0.448 (0.400) −0.537 (0.416) −0.664 (0.500) −0.407 (0.358) −0.513 (0.429) −0.303 (0.367) −0.411 (0.421) N 160 160 160 160 202 202 202 202 196 196 196 196 n umber of groups 16 16 16 16 22 22 22 22 21 21 21 21 p ( lr test v s. simple log istic re gre s-sion) 0.003 0.016 0.066 0.139 0.005 0.025 0.033 0.082 0.009 0.038 0.008 0.029
changes because they are directly applicable. New legislation might lead to more changes because of a lack of information regarding the subject of the regulation when the proposal is drafted by the Commission. Due to multicollinearity, the number of member state gov-ernments backing a proposal could not be included in the analysis as a control variable. Proposals that are backed by a blocking minority are more likely to be adopted across all stages of decision-making and for both types of success (at least partial, full) even when controlling for these factors. This effect is statistically significant at the 0.01-level.
The multilevel logistic regression analyses also show that the success probability (both full and partial success) increases with the number of votes backing a request (H2). This is true regardless of the stage of decision-making. The relationship between the number of votes
and the success probability is positive and statistically significant at the 0.01-level.2 The other
variables do not have a strong or consistent effect. Differences in success probability between different types of requests are relatively small and also not consistent across different stages.
Figure 2 presents the marginal effect of the number of votes of the requesting countries on
the predicted probability of success for the different stages of decision-making and degrees of success. The probabilities are estimated for a request for a derogation regarding a new directive with the number of recitals set to its mean value. The vertical line denotes the number of votes necessary for a blocking minority. The graph shows the positive relationship between the number of votes backing a proposal and its success probability in all of the different settings. Getting at least some concessions is a lot more likely than full success and there are only very small differences across the different stages of the legislative procedure.
The probability of success of requests is relatively large. Overall, more than a quarter of requests are completely adopted and close to half are at least partially successful. The chances of success increase with the number of votes backing a request (H2). In particular, requests which are backed by a blocking minority are more likely to be (at least partially) adopted than requests which are supported by fewer votes (H1). Thus, proposals by the Commission are frequently adapted to account for the interest of member states. These requests are not just successful within the Council but also affect the final outcome of the legislative process.
Figure 2. predicted probabilities of success rate (multilevel model).
Conclusion
Voting weights are a key ingredient in many explanations of legislative decision-making in the Council. According to rational choice theory, more votes should, ceteris paribus, translate into more influence over legislative outcomes. This study looked at the impact of voting weights on the success probability of member state requests, distinguishing between the various stages of the legislative procedure. Member states frequently put forward requests to limit the scope of European legislation, extend transitional time periods or lower standards. Close to half of the unique requests are at least partially successful and more than a quarter are fully incorporated into the final act. There is a clear difference in success rates depending on whether or not the countries requesting a change constitute a blocking minority in terms of voting weights. Varying-intercept models controlling for type of request, type of legislation (directive or regulation), amending (yes/no), salience and complexity confirm this finding. They also show that the success probability of member state requests increases with the number of votes backing a proposal. Changes incorporated in the Council’s Common Position nearly always make it into the final act. Thus, voting power in the Council does have an influence on the whole legislative process.
The study is based on all legislative procedures on environmental policy discussed in the Council between the first round of Eastern enlargement and the entry into force of the Lisbon treaty. There are two caveats related to missing information to keep in mind. First, documents for some proposals are censored, possibly due to a high level of political controversy. High political controversy should lead to hard bargaining and more changes due to member state requests, which would be in line with the findings presented here. Second, some member state requests are already incorporated in the presidency compromise proposals, which form the basis of discussions in the Council. This points to the effects of more informal negotiations in the Council. In addition, member states might anticipate resistance and not even request certain changes. Alternatively, member states might put forward requests to appease domestic constituents without seriously pursuing them. Thus, the success rate of requests cannot be equated directly with the overall bargaining success of a member state in the Council.
Previous empirical studies of member state interventions and bargaining success either found no or a negative effect of voting power, even when controlling for the relative position of member states and the salience they attach to an issue. In contrast, this study shows that there is a positive relationship between the number of votes backing a member state request and its success probability. As just pointed out, the success rate of requests is not necessarily identical to bargaining success. In addition, previous studies have compared member states individually while this study is effectively looking at member state coalitions supporting one proposal. Furthermore, the studies are using different control variables and are based on different sets of observations. Previous studies based on expert surveys are restricted to salient and controversial proposals. In contrast, the analysis presented here also includes proposals that are more typical for day-to-day decision-making in Brussels. In addition, pre-vious studies controlled for the relative position and the salience attached to an issue by individual member state governments whereas this study only includes a measure of the overall importance of a proposal. Nevertheless, the discrepancies of the empirical result merit further research.
Future research could also address the origins of member state requests (e.g. the influence of interest groups and party politics) and study the difference in the success rate between member states. Comparative case studies could also disentangle the effect of the number of member states making a request and the total number of votes (possibly by a few large member states). This might also allow us to address the issue of power versus luck by con-trolling for the position of individual member states. It would also be interesting to look at the relationship between member state requests and their voting behavior. Finally, this study is restricted to one policy field which is mainly regulatory in nature. In a regulatory domain like environmental policy (or social policy), member states in favour of higher levels of reg-ulation would acquiesce to changes that lead to lower standards because they value having a regulations at the European level. In contrast, member states that want to lower the level of regulations can credibly threaten to veto European legislation. In (re-)distristibutive policy fields this might not be the case. Similarly, member states might act differently when nego-tiating product standards for the Single Market, because domestic industries in all countries might pressure them to adopt (or prevent) specific elements. Further research is needed to study whether the findings of this study generalise to other policy fields with different char-acteristics and other time periods.
Notes
1. The data-set is available via Harvard’s Dataverse (https://dataverse.harvard.edu/dataverse/ ENLaEX). For more information, detailed examples and some analysis at the level of individual requests consult Warntjen (2017) and the web appendix provided in the supplementary material.
2. The web appendix presents the results of a series of model specifications and the complete set of regression diagnostics. The sum of votes of requesting countries has a robust positive effect on the probability of success. The appendix, the data-set and replication code is accessible at
https://dataverse.harvard.edu/dataverse/VDM. Acknowledgements
A previous version was presented at the 45th Annual Conference, University Association for Contemporary European Studies (UACES), Bilbao, September 2015, and at a workshop at NIAS in December 2015. I am grateful for helpful comments by the participants, in particular to Bernard Steunenberg and Madeleine Hosli.
Disclosure statement
No potential conflict of interest was reported by the author. References
Aksoy, Deniz. 2012. “Institutional Arrangements and Logrolling: Evidence from the European Union.” American Journal of Political Science 56: 538–552.
Arregui, Javier. 2016. Determinants of Bargaining Satisfaction across Policy Domains in the European Union Council of Ministers. Journal of Common Market Studies, online first.
Arregui, Javier, and Robert Thomson. 2009. “States Bargaining Success in the European Union.” Journal of European Public Policy 16: 655–676.
Bueno de Mesquita, Bruce, and Frans Stokman. 1994. European Community Decision-Making. New Haven, CT: Yale University Press.
European Commission. 2005. Proposal for a Directive of the European Parliament and the Council on Ambient Air Quality and Cleaner Air for Europe, COM(2005/447). Brussels: European Commission. European Commission. 2006. Proposal for a Directive of the European Parliament and of the Council
Amending Directive 2003/87/EC So as to Include Aviation Activities in the Scheme for Greenhouse Gas Emission Allowance Trading within the Community, COM(2006/818). Brussels: European Commission. Cross, James. 2012. “Striking a Pose: Transparency and Position Taking in the Council of the European
Union.” European Journal of Political Research 52: 291–315.
Cross, James. 2013. “Everyones a Winner(Almost): Bargaining Success in the Council of Ministers of the European Union.” European Union Politics 14: 70–94.
Felsenthal, Daniel, and Moshe Machover. 1988. The Measurement of Voting Power: Theory and Practice, Problems and Paradoxes. Cheltenham: Elgar.
Felsenthal, Daniel, and Moshe Machover. 1997. “The Weighted Voting Rule in the EU’s Council of Ministers, 1958–95: Intentions and Outcomes.” Electoral Studies 16: 33–47.
Felsenthal, Daniel, and Moshe Machover. 2001. “The Treaty of Nice and Qualified Majority Voting.” Social Choice and Welfare 18: 431–464.
Golub, Jonathan. 2012. “How the European Union Does Not Work: National Bargaining Success in the Council of Ministers.” Journal of European Public Policy 19: 1294–1315.
Häge, Frank. 2011. “The European Union Policy-Making Data Set.” European Union Politics 12: 455–477. Hayes-Renshaw, Fiona, and Helen Wallace. 2011. The Council of Ministers. Basingstoke: Palgrave. Hix, Simon. 2008. What is Wrong with the European Union and How to Fix It. London: Polity.
Hix, Simon, and Bjorn Hoyland. 2011. The Political System of the European Union. Basingstoke: Palgrave Macmillan.
Hosli, Madeleine, and Moshe Machover. 2004. “The Nice Treaty and Voting Rules in the Council: A Reply to Moberg. 2002.” Journal of Common Market Studies 43: 497–521.
Hosli, Madeleine, Mikko Mattila, and Marc Uriot. 2011. “Voting in the Council of the European Union after the 2004 Enlargement: A Comparison of Old and New Member States.” Journal of Common Market Studie 49: 1249–1270.
Koczy, Lazlo. 2012. “Beyond Lisbon: Demographic Trends and Voting Power in the European Union Council of Ministers.” Mathematical Social Science 63: 152–158.
König, Thomas, Andreas Warntjen, and Simone Burkhart. 2006. “The Laeken Convention. Consensus without Unity.” In Policy-making Processes and the European Constitution. a Comparative Study of Member States and Accession Countries, edited by T. König and S. Hug, 23–34. London: Routledge. Lewis, Jeffrey. 2005. “The Janus Face of Brussels: Socialization and Everyday Decision Making in the
European Union.” International Organization 59: 937–971.
Mattila, Mikko. 2004. “Contested Decisions: An Empirical Analysis of Voting in the European Union Council of Ministers.” European Union Politics 43: 29–50.
Mattila, Mikko, and Jan-Erik Lane. 2001. “Why Unanimity in the Council? A Roll Call Analysis of Council Voting.” European Union Politics 2: 31–52.
Moberg, Alexander. 2002. “The Nice Treaty and Voting Rules in the Council.” Journal of Common Market Studies 40: 259–282.
Moravcsik, Andrew. 2008. “The Myth of Europes Democratic Deficit.” Intereconomics 43(6): 331–340. Mueller, Dennis. 2003. Public Choice III. Cambridge: Cambridge University Press.
Naurin, Daniel. 2010. “Most Common When Least Important: Deliberation in the European Union Council of Ministers.” British Journal of Political Science 40: 31–50.
Naurin, Daniel, and Anne Rasmussen. 2011. “New External Rules, New Internal Games.” West European Politics 34: 1–17.
Norman, Peter. 2005. The Accidental Constitution. The Making of Europe’s Constitutional Treaty. Brussels: Eurocomment.
Panke, Diana. 2010. Small States in the European Union. Farnham: Ashgate.
Risse, Thomas. 2000. “‘Lets Argue‘: Communicative Action in World Politics.” International Organization 54: 1–39.
Scharpf, Fritz. 1996. “Positve and Negative Integration in the Political Economy of European Welfare States.” In Governance in the European Union, edited by G. Marks, P. C. Schmitter and W. Streeck, 15–39. London: Sage.
Schneider, Gerald, Daniel Finke, and Stefanie Bailer. 2010. “Bargaining Power in the European Union: An Evaluation of Competing Game-Theoretic Models.” Political Studies 58: 85–103.
Stratmann, Thomas. 1997. “Logrolling.” In Perspectives on Public Choice, edited by D. Mueller, 322–341. Cambridge: Cambridge University Press.
Thomson, Robert. 2011. Resolving Controversy in the European Union: Legislative Decision-Making Before and After Enlargement. Cambridge: Cambridge University Press.
Thomson, Robert, Frans Stokman, Christopher Achen, and Thomas König, eds. 2006. The European Union Decides. Cambridge: Cambridge University Press.
Tsebelis, George. 2002. Veto Players. How Political Institutions Work. Princeton, NJ: Princeton University Press.
Tsebelis, George. 2008. “Thinking about the Recent Past and the Future of the EU.” Journal of Common Market Studies 46: 265–292.
Warntjen, Andreas. 2010. “Between Bargaining and Deliberation: Decision-making in the Council of the European Union.” Journal of European Public Policy 17: 665–679.
Warntjen, Andreas. 2012. “Measuring Salience in EU Legislative Politics.” European Union Politics 13: 168–182.
Warntjen, Andreas. 2017. “Why No Gridlock? Coping with Diversity in the Council of the European Union.” Journal of Contemporary European Research 13: 963–979.
