Tsebelis G. Veto Players: How Political Institutions Work

Подождите немного. Документ загружается.

371

essentially adds some new stages to the cooperation procedure after the second reading of

legislation by Parliament. If in its second reading the Council disagrees with any of the

Parliamentary amendments, the text is referred to a conciliation committee, composed of equal

members of Council and Parliamentary representatives. If the committee comes to an agreement

it has to be approved by a simple majority in Parliament and a qualified majority in the Council

in order to become law. If there is no agreement, the initiative reverts to the Council, that may

reintroduce its previous position, “possibly with EP amendments,” by qualified majority or

unanimity (depending on the subject matter). Unless an absolute majority of the members of

Parliament disagrees, the law is adopted.

A comparison of the two procedures indicates several major differences. First,

Parliament has an absolute veto power in the codecision procedure, but needs an alliance with

the Commission or at least one member of the Council in order to have its veto sustained in the

cooperation procedure. Second, at the end of the codecision procedure it is the Council that

makes a “take it or leave it” proposal to Parliament. Third, in the codecision procedure

disagreement even over a single Parliamentary amendment triggers the conciliation procedure,

while in the cooperation procedure the Council could modify only those Parliamentary

amendments accepted by the Commission that had unanimous Council agreement (leaving the

others intact). Fourth, according to the codecision procedure, in certain areas (including culture,

and framework programs in R and D) decisions by the Council in the joint committee as well as

in the final stage can only be made by unanimity. Fifth, in the conciliation stage of the

codecision procedure the Commission is present, but its agreement is not necessary: if the EP

and the Council come to an agreement the position of the Commission is irrelevant.

372

d. The codecision II procedure

The EP intensely disliked the last steps of the cooperation I procedure: the fact that a

disagreement in the conciliation committee did not mean the rejection of a bill, but enabled the

Council to revert to its previous “common position” possibly including EP amendments. The

EP’s point was adopted in the Amsterdam treaty where it was recognized that a failure of the

conciliation committee to reach an agreement implied rejection of a bill. It is interesting to study

how the EP was able to impose its will on the governments that signed the next Treaty.

In order to make its preferences clear, the EP adopted a set of new rules after Maastricht.

One of these rules (#78) specified the reaction of the EP if there was no agreement in the

conciliation committee: “1. Where no agreement is reached on a joint text within the

Conciliation Committee, the [EP] President shall invite the Commission to withdraw its

proposal, and invite the Council not to adopt under any circumstances a position pursuant to

Article 189b(6) of the EC Treaty. Should the Council nonetheless confirm its common position,

the President of the Council shall be invited to justify the decision before the Parliament in

plenary sitting. The matter shall automatically be placed on the agenda of the last part-session to

fall within six or, if extended, eight weeks of the confirmation by the Council …3. No

amendments may be tabled to the Council text. 4. The Council text as a whole shall be the

subject of a single vote. Parliament shall vote on a motion to reject the Council text. If this

motion receives the votes of a majority of the component Members of Parliament, the [EP]

President shall declare the proposed act not adopted.”

Rule 78 was applied only once in the case of the draft directive on open network

provision in voice telephony (ONP). On another occasion that a joint text could not be agreed in

conciliation (on the draft directive on investment firms, credit institutions: capacity adequacy,

securities field in 1998), the Council decided not to re-affirm its common position.

373

Hix (forthcoming) has inferred that “This vote revealed that Rule 78, backed by the EP

leadership’s institutional preferences, was in fact a credible threat… This strategy paid off, as it

established compromise in the conciliation committee as the actual game equilibrium.” On the

basis of his analysis, Hix (forthcoming) concludes that the Amsterdam Treaty simply recognized

a de facto reality, that there was only one way to get any piece of legislation to be accepted: have

it proposed by the conciliation committee to both the Council and the EP (see also Corbett

(2001a)). However, the elimination of the last stage of the codecision I procedure was not a

trivial matter as Moravcsik and Nicolaïdis (1999) demonstrate. Until the last minute several

governments were against the elimination of the last step of this procedure.

We will discuss the difference between the two versions of codecision in the next section,

and we will look on some of the empirical implications of this analysis in the last section of this

chapter.

e. Extending codecision II and the qualified majority requirement

Finally, the Treaty of Nice extended the areas of applicability of the codecision II

procedure, but at the same time increased the qualified majority threshold. The qualified majority

required at Amsterdam (for a 15-membered EU) was 62/87 votes in the Council (q=.7126). Nice

(among other things) altered the weigh of different countries in the Council. For example, the

five largest countries

141

had 48/87(=55%) votes up to the Treaty of Amsterdam. They

142

receive

143/237(=60%). However, the most interesting change that will occur in the future is that along

with the enlargement of the EU to a 27 member union (by expansion to Eastern Europe) the

required qualified majority threshold will be 253/345 (q=.7333).

141

Germany, France, UK and Italy had 10 votes each, and Spain 8 votes.

142

Germany, France, UK and Italy have 29 votes each, and Spain 27 votes.

374

However this official increase in threshold is underestimating the situation. The actual

requirement specified in Article 205(4) of the Nice Treaty is a triple majority: Besides the

qualified majority threshold a Council decision has to be supported by a majority of the Member

States, and if a Member State requires it has to be supported by Members totaling 62% of the EU

population (Yataganas (2001)). All these restrictions raise significantly the qualified majority

threshold for Council decision-making, and as I demonstrated in the second chapter, increase

policy stability. I will analyze these modifications in detail in section 4.

3. Veto players analysis of legislative procedures

We will now study the implications of these complicated decision-making rules both in

terms of the distribution of legislative power among the three different institutions, and in terms

of the consequences on the decision-making discretion of other actors, like the European Count

of Justice. So, this section will be divided in two parts, first the legislative consequences of

decision-making procedures, and second the effects on bureaucracies and the judiciary.

a. Legislative Consequences of Decision-making Procedures.

In all legislative procedures agreement among the different institutions is sought. What

differs is who is the institutional actor in charge of formulating the possible agreement (agenda

setter) and what happens if such an agreement cannot be achieved (who has veto powers).

In order to address these questions I will focus in the last stages of each procedure. The

reason that I focus at the last stages is that any rational actor when called upon to decide will

look down the road on the possible consequences of his actions, and will make the decision that

will make him better off in subsequent stages of the game (and of course the final stage).

143

If all

the actors knew each other’s preferences and payoffs (the game theoretic term for such a

143

This is called “backwards induction” in the game theoretic literature.

375

situation is “complete information”) this way of thinking would lead to an immediate end of the

legislative game: The Commission would propose a bill which would be accepted by all other

actors. Indeed, the Commission would never make a proposal that would be ultimately rejected,

and the other players would not raise objections if they knew they would not win in a

confrontation. Given however that all these procedures unfold with moves and countermoves by

each actor, that cooperation reaches the second reading and that codecision goes to conciliating

committee meetings the most reasonable assumption to make is that the different players do not

operate in the ideal world of “complete information.” Let us look at the last stage of each one of

the procedures.

The last stage of the cooperation procedure is clear: The EP proposes a series of

amendments, the Commission incorporates all, some, or none of them into the final report it

submits to the Council. The Council accepts the Commission’s proposal by qualified majority, or

modifies it by unanimity. In other words, it is more difficult for the Council to modify a

Parliamentary proposal (provided it is accepted by the Commission) than to accept it. “This

procedure may enable EP to offer a proposal that makes a qualified majority of the Council

better off than any unanimous decision. If such a proposal exists, if

144

the EP is able to make it,

and if the Commission adopts it, then the EP has agenda setting powers. If, however, these

conditions are not met, the EP looses its agenda setting power. This is why I characterize EP's

agenda power under the cooperation procedure as conditional.” (Tsebelis 1994: 131).

INSERT FIGURE 11.1

I will provide an example of how the EP can make use of its conditional agenda setting

powers. I will start with a one dimensional policy space, and then move to a two dimensional

space where the results differ.

376

The Council is represented by seven members (so that the required qualified majority

can be approximated by five out of the seven members). The underlying dimension is

integration, so that the EP and the Commission are to the right of the country members, while the

status quo is to the left. The strategic calculations of the EP are as follows: It has to offer five

members of the Council a proposal that will make them better off than anything that the Council

can decide by unanimity (see Tsebelis 1994 and 1996).

Other researchers (Steunenberg 1994, Crombez 1995, Moser 1996) argue that even if the

EP behaves in the way described by Tsebelis (1994) in the second round, in the previous round,

one of two things would happen: either the Commission liked some of the amendments and

made them on its own, or, the EP knew that the Commission would reject its amendments, and

consequently did not offer them. Thus, the EP should not be making amendments either because

its opinion is already incorporated in the text, or because any changes would be rejected. In

Figure 11.1, the Commission and the Parliament have similar preferences, so the Commission

should start the procedure by making the proposal X. Either the EP would understand that no

improvement is possible, or that if it offered a different proposal (say its own ideal point), the

Commission would reject the amendment. Crombez (1995: 218) puts it succinctly this way:

“Proposition 3: Under the Cooperation procedure the Parliament's opportunity to amend the

Council's common position does not affect the equilibrium policy.” As a result they expect an

impotent EP, and argue that the agenda setting power is with the Commission. This is a

difference in the identity of the agenda setter.

Be that as it may, there is a second difference between my analysis and the expectations

of Crombez (1995), Steunenberg (1994) and Moser (1996). They believe that the agenda setter

will make a proposal that makes a qualified majority better than the status quo, while in my

144

The reader is reminded that 260 votes are required for a proposal.

377

analysis the proposal will make a qualified majority better than anything that can be promoted in

the Council by unanimity.

Figure 11.1 presents the differences of the two arguments. According to my argument,

the Council can unanimously modify the status quo and select anything in the SQ’Y’ area.

Consequently, if the EP offers X’, a proposal that member 3’ barely prefers over Y’, it will be

accepted by the Commission and members 3’, 4’, 5’, 6’, and 7’ of the Council. According to

Crombez (1995), Steunenberg (1994) and Moser (1996) the winning proposal will be located

close to point 4’ (it is the symmetric of SQ with respect to the pivotal member of the Council:

3’). In addition, the winning proposal will be made by the Commission, and the EP will make no

amendment, because it cannot improve its situation (any amendment to the right of the

Commission will be defeated, and any amendment to the left is less preferred).

A third direction of research tried to reconcile the two approaches. Bieber et al (1986:

791) argued that "With regard to the European EP, the Single Act is an inconsistent document:

Where it increases the EP's powers of participation in decision-making the practical effect is

either very limited or diminished because the exercise of the powers is conditional on the attitude

of the Council and the Commission." Similarly, Fitzmaurice (1988: 391) argued that "despite the

appearances of a co-decision model, the Council virtually retains the last word." Jacobs (1997: 6)

explicitly criticizes the first approach by making two arguments: first, that there is a tendency for

the Council to decide unanimously, and second, that the Commission has the tendency to “either

side with, or at least not go against the most powerful actor, the Council, in the final stages of the

procedure... even if it has supported Parliament amendments in first reading...”

145

A similar

145

Strictly speaking, Jacobs’ argument is not a refutation of Tsebelis’ thesis: Tsebelis (1994) “conditional agenda

setting” argument is predicated upon acceptance by the Commission, and absence of unanimity in the Council. He

does not make any prediction about how often these conditions will obtain. However, if these conditions are rarely

met, conditional agenda setting becomes less empirically relevant. For this reason we will examine Jacobs’ claims

378

argument is made by Corbett (2001a: 376): “Parliament’s powers under the cooperation

procedure to formulate a “take it or leave it” position towards Council depend on so many

conditions that, in practice, it does not usually apply.” Finally, Lodge (1987: 23) claims that the

EP’s limited power stems from the threat to block decisions: "in an alliance with one or more

member states prepared to thwart the attainment of the necessary majorities (qualified or

unanimous) unless EP's views and amendments were accommodated." Tsebelis (1996) addressed

the criticisms of the other two approaches by arguing that some of the no-impact or limited

impact theories are based on unrealistic assumptions of complete information and single

dimensionality of the issue space. Once these assumptions are relaxed, successful EP

amendments are possible. Tsebelis enumerates several such examples. For instance, the

Commission may be willing to compromise with EP because it acts as an 'honest broker', or to

avoid friction. But, because of incomplete information about EPs preferences, the Commission

may wait to observe the degree of EPs resolve first (indicated, for example, by a strong majority

or the assignment of a highly competent rapporteur) and then adopt these amendments. Tsebelis

also argues that the Commission may adopt EP amendments if the latter are introduced in a new

dimension. The lower half of Figure 11.1 indicates how the introduction of a new dimension may

make the Commission better off than its own initial proposal.

Suppose that the Commission started with a one dimensional representation of the

problem, and made the proposal X. The EP can now introduce an amendment on a different

dimension, and generate a two-dimensional policy space. In this space, the preferences of the

Council are presented by the numbers 1-7, while the EP and Commission (C) positions are

indicated to the right. Again, the EP has to make the following calculations: find what the

Council can do unanimously (anything in the area U(SQ)), and make a proposal that makes five

empirically below.

379

members better off than anything inside the area U(SQ). This is denoted by the heavily shaded

area Q(U(SQ)) in the figure. Out of all the points in Q(U(SQ)), the Parliament selects the point

closest to its own position, while the Commission has the choice of modifying this amendment

slightly to Z or (if transaction costs are high) accepting it exactly the way it was written by the

Parliament. Comparing the top and bottom of Figure 1 indicates that the Commission prefers

the proposal in a two-dimensional space (whether X or Z) to its own one-dimensional proposal.

So, the EP by introducing an amendment in a different dimension generates a different

strategic situation both for the Commission and the Council. For the Council, the new situation is

generated by the difference in opinion between members 2 and 3 along the new dimension.

Because 2 and 3 have significant differences in the two-dimensional picture (but not in the one-

dimensional projection) member 3 is willing to make many more concessions to the Commission

and the Parliament than before. The Commission prefers the Parliament’s amendment because it

makes the Commission significantly better off compared to its initial proposal. As a result,

Parliament does make a proposal and its proposal is accepted (unlike the results in single

dimensional complete information models).

Figure 11.1 simplifies the decision-making procedure in a significant way. It assumes that

the EP is a unified player, which for a Parliament representing fifteen nationalities and twelve

ideologies is a heroic simplification indeed. However, as Chapter 2 has demonstrated the results

do not change significantly if one replaces the false unified player assumption with the more

realistic “cooperative decision-making” assumption. Tsebelis (1995c) makes the case that

because of the organization of the EP, and in particular because of the “rapporteurs” of the

different bills, cooperative decision-making is a reasonable assumption.

380

This is the second time that we find concrete cases where multidimensional models lead

to different results rather than single dimensional ones, and where the case has to be made that

decision-making follows certain rules for the conclusions to follow.

146

These examples make a

strong case in favor of the veto players theory, that is based on multidimensional policy spaces.

We will see whether empirical evidence corroborates these expectations in the next section.

The Codecision I procedure has two major differences. First, it eliminates the

Commission from the last round of negotiations, and as a result reduces the influence of the

Commission on legislation. The legislative influence of the Commission is not eliminated

altogether as some researchers have argued (Crombez ***). The reason is that the Commission

can make a proposal in the beginning of the process that will be accepted by the other two actors

(the reader is reminded that the legislative process starts with a Commission proposal, so the

Commission has agenda setting role).

Second, there are two possible endings of Codecision I, depending on the outcome of the

negotiations in the conciliation committee: 1) If the conciliation committee comes to an

agreement, this agreement is introduced to both the Council and the EP; it is adopted if it

receives a qualified majority in the Council, and an absolute majority in the EP; it fails

otherwise. 2) If the conciliation committee fails to reach an agreement, the Council can make a

proposal to the EP; this proposal is considered accepted unless an absolute majority of the EP

votes against it, in which case it fails. The content of the proposal is the position “to which it

agreed before the conciliation procedure was initiated, possibly with amendments proposed by

the EP” (article 189b(6) of the Maastricht Treaty. Depending on which process is selected, the

identity of the agenda setter differs. In the first case it is the conciliation committee itself (i.e. a

combination of the Council and the EP), while in the second it is the Council alone.

146

The other was in the discussion on “pivotal” or “supermajoritarian” politics in Chapter 6.