http://eiop.or.at/eiop/texte/2000-001a.htm
: 18.1.2000| This paper's comments page | Send your comment! to this paper |
| Institutional Games Rational Actors Play – The empowering of the European Parliament(*) |
| Anders Rasmussen |
| European Integration online Papers (EIoP) Vol. 4
(2000) N° 1; http://eiop.or.at/eiop/texte/2000-001a.htm |
| Date of Publication in
: 18.1.2000 |
| | Abstract |
Back to homepage |
PDF | PS |
| This paper's comments page | Send your comment! to this paper | |
1
‘A distinguishing feature of the EP is that it does not regard itself as part of a finished institutional system, but as part of one requiring evolution or even transformation into something different.’ (Corbett, 1998:368)
This paper will make explicit a theoretical framework to explain the delegation of power to the European Parliament (EP), focusing on the change in the legislative procedures(1) that regulate its interaction with the Commission and the Council.(2) In other words, it is an analysis of the institutional consequences from an institutional(3) design, based on the assumption that different procedures result in different strategies of the actors and different outcomes of their interactions (Tsebelis, 1999).(4)
A short historic overview over the procedures in question can illustrate the EP’s increased legislative clout. In the 1957 Treaty of Rome the consultation procedure specified that the Commission enjoys a monopoly to initiate legislation, the Council accepts the proposal with qualified majority or rejects it with unanimity and the EP has to be consulted by the Council for a non-binding opinion only.
In three consecutive treaty reforms this institutional balance has been changed with the coming into force of new legislative procedures. The co-operation procedure was introduced in 1987 in the Single European Act (article 189C(**)), the co-decision procedure was introduced in 1993 in the Maastricht Treaty (article 189B), and modified and extended in 1999 in the Amsterdam Treaty (article 251).
The linearity of the progress in the powers of the EP has been the object of debate (see Tsebelis, 1996; 1997 and Moser, 1996; Scully, 1997). However, it has not been disputed that these three procedures vis-à-vis the consultation procedure give the EP a significant legislative role not previously enjoyed (Kreppel, 1997) and that the legislative power of the Commission has consistently declined (Tsebelis & Garrett, forthcoming, c).
This paper introduces a model based on rational choice institutionalism and spatial theory. The delegation of power to the EP is analysed as a long-term institutional game played by the EP, Commission, and Council. A dual analytical focus is applied, where the day-to-day policy-making is fundamental to the understanding of treaty reforms (Tsebelis & Garrett, forthcoming, a:2). It will be shown that legislative procedures have an inbuilt discretionary space to be exploited by these rational actors.
The analysis will proceed as follows. In section 2 a theoretical framework will be developed using rational choice institutionalism and spatial theory. This model will be applied to the consultation, co-operation, co-decision1 and co-decision2 procedures in section 3. Finally, section 4 will sum up the conclusions and evaluate the future applicability and validity of the framework.
2
In this paper delegation of power is analysed as a long-term institutional game played by the EP, Commission and Council.(5) The model operates on two levels, a treaty reform level and the level of day-to-day politics, illustrated in figure 1.
Initially member-states choose in a founding treaty to delegate power to supranational actors (the Commission, EP and ECJ) circumscribed by the use of legislative procedures. Assuming that the member-states have short-term horizons and imperfect information (Pierson, 1996:135ff) these procedures constitute incomplete contracts with inbuilt discretionary space for rule interpretation.(6) The rational and utility-maximising supranational actors, the legislating EP and the implementing Commission, will exploit this space to move policy outcomes closer to their ideal point by adapting strategically to the procedures. Thus the period between treaty reforms is not one of ‘consolidation’ (Moravcsik, 1993:473) but one of strategic interaction to set the precedent for the functioning of the procedures to be institutionalised by the member-states in the next treaty reform.
The analysis of the accretion of power by the EP in three treaty revisions will invariably necessitate a trade-off between analytical clarity and empirical aspirations. Given the format of this paper a comprehensive empirical account is not intended. Instead the theoretical framework is designed to reveal the causal mechanisms in the inter-institutional interaction and applying this insight to case studies illustratively.
Above it was assumed that the EP, Commission and Council are egoistic and rational actors that maximise their utility (Scharpf, 1997:20ff), yet the content of their utility-function must be specified according to the actors involved. Neglecting the EP’s oversight powers and the Council’s executive power for analytical clarity the complex array of powers(7) is simplified to a policy-making (EP, Commission, and Council) and an implementation dimension (Commission). It is assumed that the actors will want to maximise the control over outcomes with the powers invested in them. Below the two dimensions of power will be made explicit parts of the utility functions with a view to the legislative game.
In order to employ the spatial framework restrictive assumptions about the actors will be made. These are expressed in graphical form in figure 2, and its component parts explained in the sections below.
Briefly introduced, the figure features the three actors in a two-dimensional policy space of Integration and Ideology. The pivotal member of each collective actor is marked with a [· ] and dotted indifference curves are drawn for the collective actor. The shaded area designates the Pareto set within which agreement between the three actors can be found. The EP and Commission are assumed to have identical preferences on the Integration dimension, whereas the Commission and Council have identical preferences on the Ideological dimension, directing attention to potential areas of conflict.
The validity of assumptions can arguably be seen as a test of the validity of the explanations of the theory. Put more precisely, ‘the assumptions of a theory are (…) also conclusions of the theory’ (Tsebelis, 1990:32). For this reason the assumptions behind this model are not seen as stylised facts, but will be substantiated in the sections below.
3
Adhering to methodological individualism implies that the unit of analysis is an individual actor. This analytical concept needs justification when confronted with the empirical realities of the EU. From 15 nationalities, the Commission is composed of 20 Commissioners, the European Parliament’s 626 members is divided into 10 parliamentary groupings and the Council is made up of member-states of different ideological standings (Hix, 1999a:5ff).
However, Scharpf (1997:52f) has demonstrated that a collective actor with relatively homogenous preferences can be regarded as legitimate to analyse as an individual unit. Explicitly, the existence of ‘a common frame of reference that is constituted by institutional rules’ (ibid:39) allows the analysis to treat an aggregate of individuals as a unitary actor. Using this theoretical grounding the three actors will be analysed.
The complexities of the Council are reduced by the assumption of a 7 member Council with equal voting weights, a simplifying assumption shared in both power index analysis (e.g. Hosli, 1996) and in spatial theory (e.g. Tsebelis, 1994). This simplification does, however, capture the intra-Council bargaining dynamic and takes into account policy preferences by assuming majorities are made using minimum winning connected coalitions.(8)
The Commission can be divided into a political and a bureaucratic level. The College of Commissioners represents the political level, developing medium-term strategies for the EU and drafting and arbitrating in the legislative process (Hix, 1999a:32). Internally the College is driven by a strong consensus-oriented decision-making style, although votes can be taken on the request of any Commissioner. Externally the Commission is bound by the principle of common responsibility, lending credibility to the assumption of analysing a collective actor as a unitary actor. At the bureaucratic level the Commission is divided into DGs with sectoral areas of expertise, which makes it possible to identify bureaucratic entities with strongly held and shared policy-preferences (ibid:37).
Finally, treating the European Parliament as a unitary actor has perhaps been the object of most criticism (Bowler & Farrell, 1995:220). However, the assumption of a unitary actor can be justified by the EP’s committee and rapporteur system, which ‘approximates co-operative decision-making’ (Tsebelis, 1995:88). The EP’s powers of delay under the consultation procedure and amendments to legislative acts in the other legislative procedures are exercised within the committee system. In order to make the EP a more qualified legislative counterweight to the Council and Commission the political groups appoint members to the committees on the basis of their interest-group affiliation and occupational background (Bowler & Farrell, 1995:227). Rule 152 and Annex VI of the EP’s Rules of Procedure regulate the sectoral specialisation of the committees, and ensure that they in composition reflect the nationality and ideology of the aggregate EP (European Parliament, 1999).
Furthermore, in amending acts the relevant committee selects a rapporteur for the preparing of draft texts that ‘can get the support of the committee, (…) clear the floor with the required majority, and (…) get adopted by the Commission and Council’ (Tsebelis, 1995:87). Thus the rapporteur system highlights the fact that the committee system has a dual purpose of generating legislative expertise via specialisation and building coalitions via co-operative decision-making.
In summary, the three collective actors are treated as unitary actors for empirically as well as theoretically founded reasons.
4
To enhance analytical clarity most analyses apply the assumption of a one-dimensional policy space of more or less integration (e.g. Crombez, 1996; Moser, 1996; Scully, 1997; Tsebelis, 1994). In one dimension with single-peaked preferences and majority rule the median will always be a stable position, whereas the existence of a unique median in two or more dimensions, a Condorcet winner, is extremely unlikely (Hinich & Munger, 1997:62f).
However, one solution is Shepsle’s (1989) notion of structure-induced equilibrium, where institutional rules generate stable outcomes. Most legislative processes in the EU are governed by germaneness rules(9), decomposing multiple legislative dimensions into separate dimensions. The sectoral Councils of Ministers (e.g. the General Affairs, the Agriculture and EcoFin Council), the EP’s specialisation in policy specific committees and the Commission‘s DG structure are institutional equivalents of this germaneness rule. Consequently incorporating two dimensions need not result in chaos. If the dimensions can be dealt with according to the germaneness rule the added complexity may be a fair trade-off for added insight.
This paper will primarily employ the Integration dimension as it is well established theoretically and empirically. Furthermore, Hix (1994, 1998) has demonstrated that the classic left-right Ideological dimension has come to be a pervasive dimension in the EU along with the Integration dimension (see also Pollack 1999b). Integration and Ideology are assumed to be dominant, separable and non-complementary dimensions, so that the utility and preferences on one dimension can be analysed independently of the other dimension.
Incorporating Integration and Ideology in the analysis does, however, affect the institutional game. Tsebelis (1997) makes the distinction between policy (the Ideology dimension) and institutional (the Integration dimension) issues, employed in this model also. Empirically it has been demonstrated (Tsebelis et al., 1999) that conflict between the Council and the EP and Commission is most likely to occur on the Integration dimension, whereas an Ideo- logy dimension is less conflictual. How can this be theoretically sustained?
In game-theoretic terms the Integration dimension has an impact on the legislative interaction, which is directly translated into bargaining positions: Determining the level of integration and hence influence to the supranational actors (the Commission and EP) and the intergovernmental actor (Council). Thus the Integration dimension lays down the rules of the game in which the Ideology dimension is the second component of the legislative interaction. It is assumed that the Integration dimension is prior to the Ideology dimension, both in terms of long-term importance and in causality: The EP and Commission play a long-term institutional game in order to secure more favourable bargaining positions in subsequent games. Since they can be re-introduced, games on the Ideology dimension can be fought and lost without lasting effects, while a lost game on the Integration dimension through precedent will have long-term effects on the rules of the game itself.
5
A central claim in rational choice theory is that actors maximise utility. The equivalent in spatial analyses of the legislative process in the EU is the assumption that the Commission and the EP have more integrationist preferences than any member of the Council does. The existence of a vast amount of pro-integrationist amendments made by the Commission or EP to common positions of the Council verifies that the assumption empirically holds.(10) But the underlying rationale is not revealed by aggregate statistics. The integrationist preferences of the EP and Commission will therefore be theoretically substantiated below.
Ideally the MEPs are directly elected by and accountable to a European constituency, while the Commissioners are appointed by national governments, albeit not as national delegates.(11) But the agency to further integration arises from the inability of the constituency or government, respectively, to impose their ideal point ex post the election of the EP and appointment of the Commission. How is this accountability gap explained?
The EP is not constrained because European elections are second-order, fought predominantly on ‘the performance of national, with lower electoral turnout than national elections and more protest votes against the governing parties’ (Hix, 1998:50). The MEPs’ performance in the EP is thus not the benchmark for the voters’ verdict in European elections.
The independence of the Commissioners is linked to the nomination procedure. Although bureaucratic drift ideally could be minimised through ‘partisan nomination’ (Garrett, 1995:298) the Commission as a collective body is nominated and approved by all the member-states and with the Maastricht and Amsterdam Treaties subject to a vote of approval by the EP (article 214). This ‘bi-cameral’ appointment-process with veto-players of different ideological standing should ensure fulfilment of article 213, that the Commissioners exhibit ‘indisputable independence (…) and not accept instructions from any government’.(12)
The initial distinction between implementation and legislation power is one that is analytically important. Applying a principal-agent model (Pollack, 1997; 1999a), the legislating principals (EP and Council) delegate the responsibility of implementing the legislation to the agent (the Commission). The spatial location of the initial legislation and the final policy outcome may, however, diverge. This discretionary space can be traced to a number of factors.
First, the ideal policy point of the agent may not be identical with those of the principal.(13) The ability of the agent to move the policy outcome towards his own ideal point is dependent on the ease of passing new legislation (Tsebelis & Garrett, forthcoming, b:2) and how elaborate the administrative procedures regulating implementation are. If new legislation requires a unanimous decision, the discretionary space is large or if the principal has set out very detailed rules for implementation the discretionary space is small, and vice versa. The (ex ante) legislative procedures regulating the triangular interaction and the (ex post) comitology system are such examples.(14) The situation is depicted in figure 3.
6
In an institution with three legislating principals with equal voting-weight and single-peaked Euclidean preferences the median, Principal2, can introduce his ideal point, X2, which constitutes a Nash-equilibrium. The agent with ideal point Y is delegated the responsibility of implementing this act. How far can the act be skewed towards the ideal point of the implementing agent?(15) If new legislation has to be passed unanimously the agent can move the act within all of the Pareto set, and will implement the ideal point X3 of Principal3. This outcome is invulnerable to new legislation since Principal3 will veto any attempt to assert the original position of the adopted act. If new legislation can be passed with simple majority (in a three-member legislature QMV equals unanimity), the agent cannot move the policy outcome beyond X2. If X3 was sought implemented principals 1 and 2 could be made better off by introducing and passing new legislation, asserting the originally passed act at X2.
This simple example illustrates the bargaining dynamic between first, principals with different policy preferences and second, between principal and agent on the location of the implemented act. Applying this to EU legislative interaction, Principal3 and the Agent can be substituted with the EP and the Commission, respectively. It is in the interest of the Commission to implement legislative acts under a legislative procedure that gives the most integrationist actor, the EP, the most leverage. The EP and Commission can as ‘partners’ (Westlake, 1994) in a two-level game of policy-making and implementation move legislation beyond the originally adopted position and sustain this as an invulnerable position. The EP’s legislative power and the Commission’s implementation power are thus integral parts of the logic behind institutional games. Legislative power is transformed from a zero-sum game into a positive sum game, where the EP and Commission have converging interests as increasing the EP’s legislative power may increase the implementation power of the Commission.
To summarise, given that the principals (the Council and EP) have different policy preferences the institutional set-up of the legislative rules determines to what extent the agent (the Commission) can move the act towards its ideal policy points. If unanimous agreement is needed to pass new legislation the agent has maximum discretionary space (within the Pareto set), and if simple majority is needed then the agent can only move the act up till the point of the median principal. Between these two extremes, the general case is that the higher the threshold for passing new legislation, the larger the discretionary space for implementation (Tsebelis & Garrett, forthcoming, b:2).(16)
It is assumed that the actors have single-peaked Euclidean preferences arrayed as an ideal policy point, rendering the actors’ indifference curves circular with utility decreasing at the same pace as the policy outcome diverges from their ideal point. If the utility of outcome y equals z, U(y)=U(z), the actor is indifferent between z and y (Hinich & Munger, 1997:27f).
7
Analytically a distinction will be employed between day-to-day policy-making, where all three actors are involved and treaty reforms, where only the component members of the Council (the member-states) formally are involved.
The Council is modelled as seven members with equal voting weight, where a qualified majority can be simplified to 5/7th of the votes.(17) Positioning the Council on a scale from 1-7 of integration (Tsebelis & Garrett, 1997; Scully, 1997; Moser, 1997), the EP and the Commission have more integrationist preferences than any Council member does.
Legislative outcomes are analysed as a bargaining process within the Council itself and between the EP or Commission (the relevant amending actor) and the pivotal Council member: The Council is in the position of either accepting amendments with a qualified majority or rejecting amendments unanimously. The EP or the Commission will introduce the most integrationist amendment that makes the pivotal member of the Council (the member who can turn a coalition into a qualified majority) indifferent to the status quo.
The location of the status quo (SQ) or default condition itself is of importance. Ostrom (1986:12ff) points out that the default condition affects the behaviour of the negotiating actors. In a single shot legislative game there is no previous legislation to revert to should the actors fail to agree, whereas the default condition in iterated games will be the continuation of the existing policies. Thus the SQ generated under either of the legislative procedures will be to the left of the most recalcitrant member of the Council if no previous legislation has been adopted and somewhere between members 1 and 7 if previous legislation exists. To simplify the analysis the SQ is, however, assumed to be to the left of Council member 1.(18)
Acknowledging the incompatibility with rational choice and spatial theory, the insights from the structurationist work of Christiansen & Jørgensen (1999) on treaty reforms can, however, be utilised here. Their main argument is that treaty reforms instead of single events must be seen as a continuous process, structured by ‘the pre-defined demands on the IGC [and] the institutionalized allocation of interests built into the EU system’ (ibid:3). Substituting ‘institutionalized allocation of interests’ with the actors’ behaviour in day-to-day politics directs attention to the rationality behind adapting strategically to rules. The Institutional dimension is thus the most important to the EP because this dimension can be skewed towards its ideal point through strategic adaptation to the institutional rules in the process leading up to treaty reforms.
8
However, treaty reforms are, according to the liberal intergovernmentalist claim, the member-states’ finest hour: ‘European integration can best be explained as a series of rational choices made by national leaders’ (Moravcsik, 1998:18). The primacy of the member-states in treaty reforms can be linked to the aggregation rule applied, unanimous agreement. Unanimity should theoretically allow every actor to make the negotiations meet particular demands as the need for strategic alliances is eradicated by the implicit veto-option. Scharpf (1988:254) has described this phenomenon as the joint decision trap, which exposes the two distinctive traits of decision-making in treaty reforms: That the central EU governance decisions are dependent on member-state governments and that agreements must be unanimous (formally or de facto). Although this aggregation rule raises the threshold for change it also adds a ratchet-like dynamic to treaty reforms. The default condition for non-agreement is the continuation of the existing procedures, turning every member-state into a veto-player that can make the ‘de facto procedure’ invulnerable to an enforcement of the ‘de jure procedure’ (Hix, 1999b). The logic is illustrated in figure 4.
Drawing on Pierson (1996), at time T0 the member-states choose in a founding treaty to delegate power to the supranational actors, regulated by a legislative procedure. Playing the institutional game in the period between T0 and T1, these actors adapt strategically to the de jure procedure in an attempt to move it closer to their preferred outcome and working. The EP and Commission propose(19) this de facto procedure in the next treaty reform at time T1. Maintaining the assumption of Euclidean preferences, the member-states will accept proposals by the EP and Commission for reforms of the legislative procedure that make them better off or indifferent to the status quo, U(legislative procedurede facto) ³ U(legislative procedurede jure). In order to be accepted, these proposals can either be a Pareto-improvement for the member-states compared to the original procedure or an institutionalisation of the de facto procedure. The new de jure procedure at T1 is the default condition and object of continued strategic adaptation in the period until the next treaty reform at T2.
Having presented the theoretical framework, the next section will apply the model to the three treaty reforms.
Introduced in the 1957 Treaty of Rome the consultation procedure(20) starts with a Commission proposal that is submitted to the Council. Any member of the Council can introduce amendments to the proposal, but must be supported unanimously in order to be accepted. The proposal is adopted if supported by a qualified majority of the Council, and submitted to the European Parliament that issue an opinion, which the Council, however, is not obliged to follow. The act can then become law.
9
Majority voting in the Council as defined in the consultation procedure was, however, unilaterally blocked by France in 1966, insisting on the right to veto legislation that affected vital national interest (Moravcsik, 1998:193ff; Teasdale, 1993). This de facto unanimity version of the consultation procedure was acknowledged by the other member-states in the Luxembourg Compromise, analytically necessitating a distinction between the qualified majority and the unanimity version of the consultation procedure, shown in figure 5.
The unanimity version makes every member of the Council a veto-player that will reject any amendment by the Commission that differs from their ideal policy point. This generates a powerful lowest common-denominator dynamic, as member 1 will introduce his ideal point, resulting in a policy outcome at the Pareto-efficient position 1.
In contrast, under QMV the Commission will make a proposal that is preferred to the SQ by the coalition [34567]. This can be achieved if the proposal makes the pivotal member 3 indifferent between supporting the blocking coalition [12] and the integrationist coalition [4567]. Hence the Commission will propose a policy point XC at 5, or even slightly to the right of 5 as long as the distance between 3’s ideal point X3 and XC is less than the distance between X3 and the SQ, ½ XC -X3½ <½ X3-SQ½ .
In the consultation procedure the legislative influence of the EP was limited to the right of being consulted for a non-binding opinion. The Council and the Commission are not constitutionally obliged to take into account any amendments that the EP suggests.
However, from this limited formal legislative base the EP has since the first direct election in 1979 strategically tried to extract as much leverage as possible.(21) In 1980 the newly elected EP challenged the validity of a Council directive before the European Court of Justice for failing to respect its obligation to consult the EP. In the ECJ’s subsequent ‘Isoglucose ruling’ it stated that ‘due consultation of the Parliament (…) constitutes an essential formality disregard of which means that the measure concerned is void’ (ECJ, 1980).
10
The ruling’s clarification of the formal requirement of consultation and its impact on the legislative outcome is the object of only slightly varying opinions. Corbett points out that it is ‘more significant for urgent matters where any delay could cause problems that the Commission and even the Council would wish to avoid.’ (1998:120). In more stringent analytical terms Tsebelis & Money (1997:145ff) have analysed this causal link between the power of delay and its impact on legislative outcomes in terms of impatience. The basic intuition behind the (asymmetrical) bicameral model is that going through one more institution takes time and imposes additional transaction costs. Policy outcomes can thus (given the conditions above) be shifted towards the EP’s ideal point - converging with the Commission’s – under the shadow of delay. The EP adapted its 1981 Rules of Procedure to reflect this clarification. Rule 36 states that if the Commission does not take into account the EP’s amendments the Chairman or rapporteur will ‘request the proposal to be referred back to the relevant Committee for reconsideration’ (European Parliament, 1981b) thus delaying the final vote in plenary. This request for reconsideration was made automatic in rule 40 of the EP’s 1987 Rules of Procedure (European Parliament, 1987).(22) In effect, should the Commission accept the EP’s proposed amendments and incorporate it in the proposal submitted to the Council, the Council can only reject the act unanimously, thus making the EP a conditional agenda-setter (Tsebelis, 1994), albeit an informal one.(23)
Under the Luxembourg Compromise the Commission could only make a proposal that would make the most recalcitrant member of the Council better of than the SQ. This generates, as we have seen above, a strong lowest common denominator dynamic in the legislative outcome. But exactly this dynamic also gives the Commission the maximum implementation discretion to skew the policy point towards its own ideal point as illustrated in figure 6.
In other words, the Commission is able to move the outcome within the Pareto set, up to the ideal-point of the most integrationist member of the Council at position 7 that would reject new legislation to impose the original policy outcome. Moving beyond the Pareto set makes the implemented policy proposal vulnerable since the Council would be able to agree on new legislation within the Pareto set.
With the abolition of the Luxembourg Compromise (Teasdale, 1993) the Council were to revert back to the aggregation rule of QMV. As can be seen from figure 7 this clearly alters the space for bureaucratic discretion.
11
The Pareto set is now between the two pivotals, 3 and 5, since the Commission needs to accommodate the coalition of either [12345] or [34567]. The Commission would implement position 5, which 3 and 5 would support, given that ½ X5-X3½ <½ X3-XSQ½ . This, however, reduces the Commission’s implementation discretion to the Pareto set of [3-5].
But the ability to move policy points presupposes the existence of secondary legislation to implement (Tsebelis & Garrett, forthcoming, b:18). The veto option built into the Luxembourg Compromise had a systemic negative impact on the amount of secondary legislation to implement.(25) The end of this period was, arguably, marked by the introduction of the Commission’s 1985 White Paper to establish a Single European Market, containing 279 directives to be implemented for a 1992 deadline (Moravcsik, 1991:19).
Adopting these proposals under the unanimity of the de facto consultation procedure designates a Pareto efficient welfare-optimum. But confronting this optimistic ideal with discrepancies such as transaction costs in multiple veto-actor arenas and strategic behaviour (e.g. misrepresenting preferences or delaying final votes) it is difficult to sustain the welfare-optimal claims. Hence the Commission’s White Paper was accompanied by a package of ‘procedural reforms designed to streamline decision making’ (ibid:20).
Faced with a collective incentive to implement the Commission’s White Paper (Moravcsik, 1991) and the EP’s power of delay, the member-states accepted the Commission’s recommendations. The Single European Act accordingly introduced the co-operation procedure (article 189C), which came into force on 1 July 1987. The procedure starts with a Commission proposal that is submitted to the EP and the Council. In its first reading the EP states its opinion to the Council that acting by qualified majority adopts a common position.
In its second reading the EP has three months to choose its course of action:
The de facto working of the EP’s agenda-setting role under the consultation procedure can thus be seen institutionalised in the EP’s capability as formal conditional agenda-setter.
In submitting its proposal for the EP and the Council, the Commission will want to make the most integrationist proposal that is preferred to the SQ by a qualified majority in the Council, the coalition [34567], making member 3 pivotal. This is illustrated in the figure below.
12
In the second reading the Commission as the initial agenda-setter is replaced by the EP as the ‘conditional agenda-setter’ (Tsebelis, 1994). Positively, ‘if the EP manages to make a proposal that make the Commission and a qualified majority of the Council better off than legislation that can be voted unanimously’ (ibid:131). If such a proposal exists(27) the Commission’s gatekeeper function is merely a formal one. Negatively, if the EP rejects the Council’s common position as only a unanimous Council can override this rejection.
The policy outcome does not differ from the (hypothetical) QMV version of the consultation procedure. The EP will make the proposal that makes the pivotal member 3 indifferent between the blocking coalition [12] and the integrationist coalition [4567]. Hence the EP will propose a policy point XEP at 5, or even slightly to the right of 5 as long as the distance between 3’s ideal point X3 and the EP’s proposed ideal point XEP is less than the distance between X3 and the SQ, ½ XEP -X3½ <½ X3-XSQ½ .
Building on a report on the democratic deficit arising from the institutional imbalance between the Council and the EP (European Parliament, 1988a) the EP stated its intention to ‘exploit to the maximum the opportunities offered by the Single Act’ (European Parliament, 1988b:7). Accordingly the EP set out to explore the potentials of the co-operation procedure.
The power of delay in the consultation procedure is a characteristic retained in the EP’s first reading of the co-operation procedure as no time limits exist for deliberation. The act can thus effectively be aborted if the EP refers the act back to the relevant committee for sustained deliberation. Should the Commission not take into account the EP’s amendments rule 40 of the Rules of Procedure makes this an automatic step. Further, rule 36(3) commits the relevant committee to ascertain that the correct legal basis and consequently the correct procedure has been applied for the Commission’s proposal before considering the act (European Parliament, 1987; 1988b:14).
In its second reading, the EP furthermore designed its Rules of Procedure to reflect that the co-operation procedure entails a veto-option and a fixed deadline for deliberations. If the EP rejects the Council’s common position, the act will either fall or can be overruled by a unanimous Council. Before voting in plenary for a rejection of the common position the Commission is according to rule 50(2) and 51(4) asked to announce its opinion on the EP’s amendments (European Parliament, 1987). The EP’s conditional agenda-setting is thus made less conditional, since the Commission’s announcement of its opinion on the EP’s amendments under the shadow of delay is skewed towards acceptance rather than rejection of the EP’s proposals. The second reading’s fixed deadline further weakens the bargaining position of the Council, since it only has three months to find the unanimity required to reject the EP’s amendments (Corbett, 1998:264).
The aggregate statistics available confirm that the EP was successful in having amendments adopted. The proportions of amendments accepted in the first and second reading respectively were by the Commission 54 % and 43 % and by the Council 41 % and 21 % (European Parliament, 1997, Annex 2:6). It should be remembered that quantitative statistics do not reveal whether the amendments were politically important or mere technicalities. The qualitative dimension of the bargaining dynamic can best be illustrated with an empirical example, the Car Emission Directive, where all the legislative provisions were employed.
13
On 15 February 1988 a proposal for reducing the emission from small cars was submitted to the Council and EP. In its first reading the EP adopted amendments calling for the introduction of more stringent standards than envisaged by the Commission, which consequently rejected the amendments. The Council, however, adopted the Commission’s proposal by QMV and submitted this common position to the EP (Tsebelis, 1994:136).
In its second reading the EP’s committee on environment, backed by rule 50(2) and 51(4) (European Parliament, 1987), stated its intention to reject the common position should the Commission not accept its amendments. Hence the Commission accepted to ‘incorporate the amendments into a revised proposal’ (Corbett, 1998:266), which it presented to the Council. The Council could accept this proposal with QMV or amend it unanimously. At least three Council members (Denmark, Holland and Germany) threatened to introduce stricter standards unilaterally (Hubschmid & Moser, 1997:240), thus undermining the potential for unanimous action. Consequently, the Council adopted the EP’s amended proposal by QMV.
As shown in figure 9, the Commission’s bureaucratic discretion under the co-operation procedure is the same as under the QMV version of the consultation procedure, as long as the EP and Commission are more integrationist than any member of the Council is.
The 5/7th majority rule makes members 3 and 5 pivotal, since the EP’s amendment will need a coalition of either [12345] or [34567]. The EP would propose member 5’s ideal point, which member 3 would support, given that ½ X5-X3½ <½ X3-XSQ½ . The Commission’s implementation discretion is thus confined to the Pareto set of [3-5].
The member-states introduced the co-decision1 procedure (article 189B) in the Maastricht Treaty, which came into effect on 1 November 1993. It starts with a Commission proposal that is submitted to the EP and the Council. The EP states its opinion in its first reading to the Council that acting by qualified majority adopts a common position.
In its second reading the EP has three months to choose its course of action:
14
It follows that disagreement between the EP and Council in the second reading will lead to the convening of the Conciliation Committee (CC), the most significant difference from the co-operation procedure. The EP is consequently enabled to re-engage in the legislative process if it rejects the Council’s common position and the Council does not let the act fall or if the Council does not accept its amendments.
The CC is composed of two delegations from the Council and the EP, with the Commission acting as mediator (article 189B(4)). It is given 6 weeks to produce a joint text, which first has to be approved by QMV for the Council’s delegation and simple majority for the EP’s delegation and then with absolute majority by the EP and with QMV by the Council.
If a joint text cannot be found the Council can within six weeks confirm by QMV its common position in order to adopt the act, while the EP has six weeks to reject the act by an absolute majority of its component members, thus giving it an unconditional veto-power.
Legislative outcomes under the co-decision procedure are modelled as a bargaining process within the Council itself and between the EP and the pivotal Council member. ‘(G)iven that rational actors use backward induction in their decision-making, any subgame perfect equilibrium solution of the whole procedure would prescribe equilibrium behavior in each subgame’ (Tsebelis, 1997:34). In other words, the Council will not settle for a policy outcome that makes it worse off than what could be achieved in the Conciliation Committee. In this endgame a joint text needs to be supported by a qualified majority in the Council, hence member 3 is pivotal and makes a take-it-or-leave-it proposal to the EP assuming that it prefers any proposal to the SQ (Garrett, 1995). According to Tsebelis & Garrett (1996) this proposal XCouncil will be a reflection of what the coalition [34567] prefers to the status quo, hence member 3 ‘will be able to impose its will on the Council’ (ibid:289) and propose its own ideal point 3. Knowing that the Council can confirm its common position, the EP cannot be made better of by vetoing and will accept X3.
15
This analysis of the endgame is disputed by Scully (1997) who argues that the policy outcome will be a reflection of the Commission’s initial proposal. Given that the Commission sides with the EP on the Integration dimension, it will make a proposal XC slightly to the left of member 6 (or at least on the continuum of members 4,5,6), which the EP would accept without amending and which the Council would adopt by QMV [34567]. Empirically, however, this argument cannot account for the EP’s amendments to the Commission’s proposal (Kreppel, 1997). This paper will consequently retain the Tsebelis & Garrett assumption that the policy outcome will be at position, shown in figure 10.
The EP had argued for co-decision ‘with the two institutions on an equal footing’ (European Parliament, 1994f), but the agenda-setting ability of the Council to make a take-it-or-leave-it proposal in the endgame creates a procedural imbalance. Adapting to tackle this potential problem and to take effect of the co-decision1 procedure the EP designed its 1994 Rules of Procedure, rule 78, to automatically put a confirmed Council common position to a vote for rejection (European Parliament, 1994c).
According to the latest Progress Report (European Parliament, 1998b:22) out of a total of 130 completed co-decision procedures at 31 July 1998 only three (2.3 %) resulted in failed conciliation. To move beyond these statistics, a detailed qualitative analysis of the only directive putting rule 78 to the test, the directive on Open Network Provision (ONP) in July 1994 (EP, 1994c) should potentially reveal the strategic features and provisions of the co-decision1 procedure and the inherent bargaining dynamic.
The ONP directive was on 29 March 1994 taken to the Conciliation Committee as the fourth act under the co-decision1 procedure (European Parliament, 1997a, Annex II:8) since the Commission accepted only four out of a total of 14 of the EP’s amendments to the Council’s common position in its second reading (Commission, 1994:2).
A closer inspection of the 10 rejected amendments reveals that the Commission had a favourable opinion on the EP’s amendments to the comitology procedures (3,12,14), which it, however, was forced to reject ‘because any changes to comitology procedures require an interinstitutional agreement’ (ibid:3)(29).
This question of consultation and the underlying theme of comitology links directly to the institutional balance and the bureaucratic discretion of the Commission. Without going into detail and neglecting the safeguard procedure, three basic committees exist: The advisory committee (which only can advise), the management committee (which can block) and the regulatory committee (which the Commission shall follow). It is therefore in the interest of the Commission to be restricted as little as possible, hence ‘its preference for an advisory Committee’ (Commission, 1994:5). This committee is also preferred by the EP, since this grants the Commission the maximum of autonomy from the ‘government-appointed scientific experts’ (Garman & Hilditch, 1998:1351) and allows the EP to be involved in the committee procedure.
16
This can be seen as one of the long-term institutional goals the EP set for itself with the co-decision procedure, as reflected in its resolution of 16 December 1993 on comitology, claiming that ‘full responsibility of legislative acts (…) covered by provisions of Article 189B lies with the Council and Parliament’ (European Parliament, 1994b:176). In other words, altering the institutional balance towards recognition of the EP as a co-legislator also in the implementation of acts.
The Council’s reception of the Commission’s opinion did, however, not satisfy the EP’s aspirations. On the contrary, in the Conciliation Committee the Council was unwilling to incorporate EP amendments or produce a joint agreeable text (European Parliament, 1994e:5), and ultimately chose to confirm its common position (Council, 1994).
Comitology was consequently solely a matter of determining the position of the default condition on the Institutional dimension. By confirming its common position, the Council in effect placed the SQ to the left of its position (i.e. setting a precedent), thus limiting the EP’s choice of action to accepting this new position of the SQ or preserving the SQ by vetoing the Council’s confirmed common position. Conforming to the provisions laid down in rule 78 of its Rules of Procedure, the EP put the common position to a vote of rejection, meeting the absolute majority needed with 379 votes to 45 (European Parliament, 1995a).
This challenges the Tsebelis & Garrett (1996; 1997) assumption that the EP at all times would prefer any pro-integrationist outcome to the status quo. First, that the Council in the course of conciliation should make an offer that is only slightly better than the SQ to achieve a joint text and successful conciliation and second, knowing that the Council is able to make a take-it-or-leave-it proposal if conciliation fails, the EP should accept such an offer.
If, however, it is acknowledged that the EP considers the interactions in the Conciliation Committee as an iterated institutional game, reputational considerations have to be incorporated in the analysis (Tsebelis, 1990:156). The shadow of the future from a single interaction is then very long, making reputational investments rational. By sacrificing short terms gains on the Ideological dimension for long term gains on the Integration dimension the EP strengthen its bargaining power, thus forcing the Council to be more accommodating in its conciliation behaviour. Vetoing should therefore establish a benchmark for the EP’s willingness to be flexible and accommodating in finding compromises that are jointly agreeable.
Furthermore, it can be argued that vetoing is not only a preservation of the SQ, but indeed an improvement. Albeit an integral part of the Tsebelis & Garrett argument, in confirming its common position the Council is according to its own legal service ‘restricted in its choice of Parliament’s amendments which it wishes to take aboard to those formally adopted by the Parliament in second reading’ (Jacobs, 1997:12). Vetoing in the formative stage of co-decision(30) must therefore be seen as a strategic and rational action to maximise the EP’s influence on future negotiations.
17
The Commission’s bureaucratic discretion under the co-decision1 procedure increased compared to the co-operation procedure. Using backward induction, the Council would not suffice with a policy outcome that made it worse off than what could be achieved in the Conciliation Committee. Although the Council’s pivotal member 3 can be an agenda-setter by making a take-it-or-leave-it proposal to the EP if a joint text cannot be found, the EP has the unconditional veto power to reject this proposal.
While the policy-outcome can be identified on the continuum [3-5], according to the Scully (1997) and Tsebelis & Garrett (1997) dispute, as long as the EP and Commission are more integrationist than any member of the Council is, the EP would not support new legislation to overrule the implemented policy. Thus the Commission’s implementation discretion vis-à-vis the co-operation procedure is expanded to the set of [3(5)-EP], which from the Council’s point of view may exceed a Pareto-efficient [1-7] outcome. Any outcome within this continu-um is thus rendered invulnerable to new legislation.
Introduced in the Amsterdam Treaty, which came into effect on 1 May 1999, the modified co-decision2 procedure (article 251) is identical with co-decision1 until the convening of the conciliation committee, with two exceptions. First, the EP can declare after the first reading that it approves the act after which the Council can adopt it immediately, and second, that the EP can reject the proposal in a single vote without having to announce its intention to reject sustained by a second vote.
Should the conciliation committee fail to produce a joint text, the EP will no longer have to reject it in the third reading, as the Council cannot confirm its common position. This means that if conciliation fails the act lapses, thus making the conciliation committee the final stage in the procedure, an institutionalisation of the de facto working of the co-decision1 procedure since the EP’s rejection of the ONP directive.
Probing the empirical impact is currently impossible, but the theoretical implications of the co-decision2 procedure are apparent in the endgame, illustrated in figure 12 below.
18
The Conciliation Committee can be analysed as a non-cooperative bargaining game (Tsebelis & Garrett, forthcoming, c:25) where the policy-outcome generated will be a Nash-equilibrium ‘split-the-difference’ between the pivotal member of the Council and the EP, taking a discrete value between 5 and 6, ½ (XEP-X3)/2½ . However, assuming that the SQ is not to the left of member 1 and that one of the actors is institutionally advantaged (e.g. time-preferences) the position of the policy outcome is obviously skewed towards one of the actors (ibid:27).
As shown in figure 13, the revised co-decision procedure does, however, not change the findings of the bureaucratic discretion of the Commission from the co-decision1 procedure.
With the abolition of the third reading the EP shall not find an absolute majority to reject a confirmed common position of the Council. Implementing the ideal point of the EP is thus made invulnerable to new legislation, defining the Commission’s bureaucratic discretion to the interval [3-EP].
To understand why member-states delegate legislative power to supranational actors this paper has made the proposition that the EP, Commission and Council play an institutional game in which they adapt strategically to the legislative procedures. Paraphrasing Westlake (1994) the EP and Commission are partners in the EU policy-making process, as they have identical integrationist preferences. Conceptualising legislative rules as incomplete contracts with inbuilt space for interpretation, the EP has adapted strategically to move the policy-outcome closer to its ideal policy position, which in effect has increased the Commission’s bureaucratic discretion. The EP and Commission can therefore, in a two-level game of policy-making and implementation, move legislation beyond the originally adopted position and sustain this as an invulnerable position.
However, only under conditions of incomplete information will the initial choice of institutions result in unintended institutional consequences (Pierson, 1996). With each institutional game played, information is generated on the preferences, strategies and pay-offs of the actors, thereby reducing the discretionary space for future interpretation of the procedures.
The delegation of power to the EP, this paper has advocated, can be analysed as an institutional game played by rational actors. Acknowledging that different legislative procedures generate different strategies of the actors and different outcomes of their interactions (Tsebelis, 1999) this paper has shown the underlying rationality behind the increase in the legislative powers of the EP and the decrease of the powers of the Commission. The EP and the Commission can, in a two-level game of legislating and implementing, move legislation beyond the originally adopted position and sustain this as an invulnerable position.
19
Given the institutional set-up of unanimity in treaty reforms, member-states face the ratchet-like logic of delegating power to supranational actors. This ‘joint-decision trap’ makes it impossible to impose a legislative procedure that enforces the original intent, as all member-states are veto players. Having Euclidean preferences, the member-states will be indifferent between the de facto operation of the original de jure procedure and a new de jure procedure that institutionalises this. Consequently institutional procedures have an impact on the behaviour of bargaining member-states in treaty reforms.
The fundamental assumption in this paper is that the EP and Commission have more integrationist preferences than any member of the Council does. Although the very low turnout in the recent European elections disconfirms the thesis that giving the EP more power would make it more accountable, the long-term impact from empowering the EP is open to speculation.
Indeed, one of the more speculative scenarios would be that the Council has in fact played a very long-term institutional game under the assumption that the second-order mirror-image quality of the EP will not persist. The impact on policy-outcomes and bureaucratic discretion will be anything but negligible as other institutional features, such as the position of the SQ, suddenly will be the determining factors behind the pace of integration in Europe. Should such a scenario materialise it only serves to demonstrate one of the main points of this paper, the causality between institutional consequences and institutional choices.
Bowler, Shaun & Farrell, David (1993) ‘Legislator Shirking and Voter Monitoring: Impacts of European Parliament Electoral Systems upon Legislator-Voter Relationships’, Journal of Common Market Studies, 31:1 (45-69)
Bowler, Shaun & Farrell, David (1995) ‘The Organizing of the European Parliament: Committees, Specialization and Co-Ordination’, British Journal of Political Science, 25 (219-243)
Christiansen, Thomas & Jørgensen, Knud Erik (1999) ‘The Amsterdam Process: A Structurationist Perspective on EU Treaty Reform’, European Integration Online Papers, 3:1, http://eiop.or.at/eiop/texte/1999-001a.htm
Commission (1981) COM(81)581/fin., 14.10.1981
Commission (1994) Opinion of the Commission, (ONP, co-decision 2. reading) COM(94)48fin. COD 437
Commission (1995) Report on the operation of the Treaty on European Union, SEC(95)731/fin., 10.05.1995
Corbett, Richard (1998) The European Parliament’s Role in Closer EU Integration, London: MacMillan Press Ltd.
Crombez, Cristophe (1996) ‘Legislative Procedures in the European Community’, British Journal of Political Science, 26 (199-228)
Crombez, Christophe (1997) ‘Policy-making and Commission Appointment in the European Union’, Aussenwirtschaft, 52 (63-82)
Council (1994) Common Position as adopted on 07.02.1994, OJ C101, 09.04.1994
Council (1995) Draft Report of the Council on the Functioning of the Treaty on European Union, 5082/95
Dunleavy, Patrick (1991) Democracy, Bureaucracy & Public Choice, London: Harvester Wheatsheaf
Dunleavy, Patrick (1997) ‘Explaining the Centralization of the European Union’, Aussenwirtschaft, 52 (183-212)
European Court of Justice (1980) Isoglucose cases 138/79 and 139/79
European Parliament (1981a) Resolution, 1-207/81, 09.07.1981
European Parliament (1981b) Rules of Procedure, 2nd edition
European Parliament (1984a) Rules of Procedure, 3rd edition
European Parliament (1984b) Draft Treaty Establishing the European Union, OJ 77/1984, 14.02.1984
European Parliament (1986) Resolution, Minutes, 16.01.1986
European Parliament (1987) Rules of Procedure, 4th edition
European Parliament (1988a) Report on the democratic deficit in the European Community, PE111.236/fin.
European Parliament (1988b) Report on the results obtained from implementation of the Single Act, PE121.245/fin.
European Parliament (1994a) Voice telephony and ONP, Co-decision procedure: second reading, A3-0006/94
European Parliament (1994b) Resolution on questions of comitology, OJ C20, A3-0417/93
European Parliament (1994c) Voice telephony and ONP, Co-decision procedure: third reading, A4-0001/94
European Parliament (1994d) OEIL http://www1.europarl.eu.int/scripts/enviewproc.idc?lang=2&id=1479
European Parliament (1994e) Session Document 15 July 1994, A4-0001/94, PE209.088/fin.
European Parliament (1994f) Report on the first applications of the co-decision procedure, doc_en\253\253629
European Parliament (1995a) Progress Report: Second Half of 1994, PE211.522
European Parliament (1995b) Reflection Group’s Report, http://www.europarl.eu.int/en/agenda/igc-home/eu-doc/reflect/final.html
European Parliament (1995c) Modus vivendi of 20 December 1994, OJ C293
European Parliament (1996) Progress Report: 1 March 1995 to 31 July 1996, PE216.734
European Parliament (1997a) Progress Report: 1 August 1996 to 31 July 1997, PE223.209, http://www.europarl.eu.int/dg2/conciliation/rep96-97/en/report.htm
European Parliament (1997b) Intergovernmental Conference, Briefing no.8: The Co-decision Procedure, http://www.europarl.eu.int/dg7/fiches/en/fiche8.htm
European Parliament (1997c) Intergovernmental Conference, Briefing no.21: Comitology, PE166.676
European Parliament (1998a) Report on the new co-decision procedure after Amsterdam, PE225.915/fin.
European Parliament (1998b) Progress Report: 1 August 1997 to 31 July 1998, PE225.314, http://www.europarl.eu.int/dg2/conciliation/rep97-98/en/report.htm
European Parliament (1999) Rules of Procedure, 14th edition, http://www2.europarl.eu.int/dg7in/seid.pl?APP=REGLEMENT+PRG=SUMMARY+FILE=REGLEMENT-PE-EN+LANGUE=EN
Franchino, Fabio (1998) ‘Institutionalism and Commission’s Executive Discretion: An empirical analysis’, European Integration Online Papers, 2:6, http://eiop.or.at/eiop/texte/1998-006a.htm
Garman, J. & Hilditch, L. (1998) ‘Behind the scenes: An examination of the importance of the informal processes at work in conciliation’, Journal of European Public Policy, Vol.5:2 (271-284)
Garrett, Geoffrey (1995) ‘From the Luxembourg Compromise to Co-decision: Decision Making in the European Union’, Electoral Studies, 14:3 (289-308)
Golub, Jonathan (1997) ‘In the Shadow of the Vote? Decisionmaking Efficiency in the European Community, 1974-1995’, MPIfG Discussion Paper, 97:3 (1-48), http://www.mpi-fg-koeln.mpg.de
Hinich, Melvin J. & Munger, Michael C. (1997) Analytical Politics, Cambridge: Cambridge University Press
Hix, Simon (1994) ‘Approaches to the Study of the European Union: The Challenge to Comparative Politics’, Western European Politics, 17:1 (1-30)
Hix, Simon (1998) ‘Approaches to the Study of the European Union II: The ‘New Governance’ Agenda and its Rivals’, Journal of European Public Policy, 5:1 (38-65)
Hix, Simon (1999a) The Political System of the European Union, London: MacMillan
Hix, Simon (1999b) ‘Institutional Agenda-Setting Through Discretion in Rule Interpretation’, Paper presented at Nuffield College, Oxford University, May 1999
Hosli, Madeleine (1996) ‘Coalition and Power: Effects of qualified majority voting in the Council of the European Union’, Journal of Common Market Studies, 34:2 (255-273)
Hubschmid, Claudia & Moser, Peter (1997) ‘The Co-Operation Procedure in the EU: Why was the European Parliament influential in the decision on Car Emission Standards?’, Journal of Common Market Studies, 35:2 (225-241)
Jacobs, Francis B. (1997) ‘Legislative co-decision: A real step forward?’, Paper produced for the fifth biennial ECSA Conference in Seattle, May 29 to June 1 1997
König, Thomas & Bräuninger (1997) ‘Decisiveness and Inclusiveness: Intergovernmental Choice of European Decision Rules’, European Integration Online Papers, 1:22, http://eiop.or.at/eiop/texte/1997-022.htm
Kreppel, Amie (1997) ‘The European Parliament’s Influence over EU policy outcomes: Fantasy, Fallacy or Fact?’, Paper prepared for the 5th biennial meeting of the European Community Studies Association, May 28-June 1, 1997, Seattle, Washington, USA
Moravcsik, Andrew (1991) ‘Negotiating the Single European Act: National Interests and Conventional Statecraft in the European Community’, International Organization, 45:1 (19-56)
Moravcsik, Andrew (1993) ‘Preferences and Power in the European Community’, Journal of Common Market Studies, 45:1 (473-524)
Moravcsik, Andrew (1998) The Choice for Europe, London: UCL Press
Moser, Peter (1996) ‘The European Parliament as a Conditional Agenda-Setter: What are the Conditions?’, American Political Science Review, 90:4 (834-838)
Ostrom, Elinor (1986) ‘An Agenda for the Study of an Institution’, Public Choice, 48 (3-25)
Pierson, Paul (1996) ‘The Path to European Integration: A Historical Institutionalist Analysis’, Comparative Political Studies, 29:2 (123-163)
Pollack, Mark (1997) ‘Delegation, Agency and Agenda-setting in the European Community’, International Organization, 51:1 (99-134)
Pollack, Mark (1999a) ‘Delegation, Agency and Agenda-setting in the Treaty of Amsterdam’, European Integration online Papers, 3:6, http://eiop.or.at/eiop/texte/1999-006a.htm
Pollack, Mark (1999b) ‘Beyond Left and Right? Neoliberalism and regulated Capitalism in the Treaty of Amsterdam’, University of Wisconsin Working Paper Series in European Studies, 2:2, http://polyglot.lss.wisc.edu/eur/works/pollack.html
Scharpf, Fritz (1988) ‘The Joint-Decision Trap: Lessons from German Federalism and European Integration’, Public Administration, 66:3 (239-278)
Scharpf, Fritz (1997) Games Real Actors Play, Oxford: Westview Press
Scully, R.M. (1997) ‘The European Parliament and the Co-decision procedure: A Reassessment’, Journal of Legislative Studies, Vol. 3:3 (58-73)
Shepsle, Kenneth (1989) ‘Studying Institutions: Some Lessons from the Rational Choice Approach’, Journal of Theoretical Politics, Vol. 1:2 (131-147)
Shepsle, Kenneth & Weingast, Barry R. (1981) ‘Structure-induced Eqilibrium and Legislative Choice’, Public Choice, 37 (503-519)
Teasdale, Anthony L. (1993) ‘The life and death of the Luxembourg Compromise’, Journal of Common Market Studies, 31:4 (567-579)
Tsebelis, George (1990) Nested Games: Rational Choice in Comparative Politics, Berkeley: University of California Press
Tsebelis, George (1994) ‘The Power of the EP as a Conditional Agendasetter’, American Political Science Review, 88:1 (128-142)
Tsebelis, George (1995) ‘Conditional Agenda-Setting and Decision-Making Inside the European Parliament’, Journal of Legislative Studies, 1:1 (65-93)
Tsebelis, George (1996) ‘More on the European Parliament as a Conditional Agenda-setter’, American Political Science Review, 90:4 (839-852)
Tsebelis, George (1997) ‘Maastricht and the Democratic Deficit’, Aussenwirtschaft, Vol. 52:1/2 (29-56)
Tsebelis, George (1999) ‘Institutional Analyses of the European Union’, European Community Studies Association Newsletter, Spring 1999
Tsebelis, George & Garrett, Geoffrey (1996) ‘An Institutional Critique of Intergovernmentalism’, International Organization, 50:2 (269-299)
Tsebelis, George & Garrett, Geoffrey (1997) ‘Agenda-setting, Vetoes and the European Union’s Co-decision Procedure’, Journal of Legislative Studies, Vol. 3:3 (74-92)
Tsebelis, George & Garrett, Geoffrey (forthcoming, a) ‘An Institutional Critique of Neofunctionalism’, International Organization
Tsebelis, George & Garrett, Geoffrey (forthcoming, b) ‘The Institutional Determinants of Supranationalism in the European Union’
Tsebelis, George & Garrett, Geoffrey (forthcoming, c) ‘Legislative Politics in the European Union’, European Union Politics, 1:1, 2000
Tsebelis, George & Kalandrakis, Anastassios (forthcoming) ‘European Parliament and Environmental Legislation: The case of chemicals’, European Journal of Political Research
Tsebelis, George et al. (1999) ‘Legislative Procedures in the European Union: An empirical analysis’, http://www.sscnet.ucla.edu/tsebelis
Tsebelis, George & Money, Jeannette (1997) Bicameralism, New York: Cambridge University Press
Weiler, Joseph H.H. (1991) ‘The Transformation of Europe’, Yale Law Journal, 100 (2403-2483)
Westlake, Martin (1994) The Commission and the Parliament: Partners and Rivals in the European Policy-Making Process, London: Buttersworths
(*) I would like to thank the two anonymous referees of the EIoP for helpful comments.
(**) References to Treaty articles are made according to the Treaty in force at the time of the event.
(1) Consequently the empowering of the EP in the scrutiny and investiture procedure of the Commission (article 201), the threat of which made the Commission resign in March 1999, and its budgetary powers (article 272) will not be dealt with here.
(2) The European Court of Justice is recognised as a political actor in its own right (e.g. Weiler, 1991), but for reasons of analytical clarity its impact on the legislative game is beyond this paper's format.
(3) 'Institution' is defined as formal or informal rules that are experienced by the actors as authoritative because defection will be sanctioned or at least is felt as a credible threat. See Ostrom (1986) and Scharpf (1997:38ff).
(4) This has been demonstrated empirically in recent works of Golub (1997); Kreppel (1997); Tsebelis (1999); Tsebelis et al. (1999) and Tsebelis & Kalandrakis (forthcoming).
(5) The terms [EP], [Council] and [Commission] are used as shorthand for more sectoralised bodies. The Council is composed of sectoral ministers from the incumbent governments of the member-states, the EP by specialised committees and the Commission by relevant DGs.
(6) I am grateful to Simon Hix for this argument.
(7) The division of powers is as follows (Hix, 1999a:21ff). The Commission executive (implementing policies) and legislative power (drafting legislative acts); the EP oversight (e.g. investiture and scrutiny of the Commission) and legislative power (varying according to the legislative procedures). Finally, the Council has executive (treaty reforms and setting of the long term political agenda) and legislative powers (agreement by unanimity or QMV is a precondition for the passing of any legislative act).
(8) See Tsebelis & Garrett (forthcoming, a) for a detailed discussion of the analytical implications arising from the difference between minimum winning coalition and minimum winning connected coalition.
(9) A strict definition entails that preferences are separable and that issues are voted at one at a time (Hinich & Munger, 1997:61).
(10) E.g. the EP has made 4572 amendments under the co-operation procedure in the period 1987-93 (Tsebelis & Garrett, 1997:87) and 2038 amendments under the co-decision1 procedure in the period 1993-97 (Tsebelis et al., 1999:16).
(11) A scenario where commissioners are seen as national delegates, constrained by the appointing member-states has been advanced by Crombez (1997).
(12) Despite this ideal vision, Hix points out that 'commissioners tend to have significant links to partisan domestic constituencies' (1999a:35).
(13) Public choice theory lists a number of factors that may generate bureaucratic drift, such as, capture by interest groups, and bureau-shaping (Dunleavy) or budget-maximising (Niskanen) strategies.
(14) The EP and Commission have vehemently opposed the comitology system (European Parliament, 1994b; 1997c), seeking instead to grant the Commission a maximum of autonomy and to allow the EP involvement in the Committee procedure.
(15) See Tsebelis & Garrett (forthcoming, a:15ff) and Hix (1999a:23ff) for a more detailed account.
(16) Franchino, however, argues reversely: '(T)he executive discretion of the Commission is likely to be larger under qualified majority vote rather than unanimity' (1998:6).
(17) Regulated by article 205 the threshold for a qualified majority in the Council is currently 62/87=0.7126 where 5/7= 0.7143.
(18) The position of the SQ is of particular importance when it is between Council member 1 and 7 if the Council collectively should want to 'roll back integration', making the EP's veto-power more important than conditional agenda-setting power. See Tsebelis (1997:48ff) for a detailed discussion.
(19) The Commission has always been present as a non-voting actor to facilitate negotiations. In the Maastricht and Amsterdam Treaty negotiations the EP was allowed to have an 'IGC-task force' present too.
(20) The descriptive part is based on (European Parliament, 1988b).
(21) The perhaps most spectacular account of agenda-setting by the EP was its 1984 draft Treaty of European Union (European Parliament, 1984b). Key points concerned the strengthening of the Commissions executive discretion and enhancing the democratic legitimacy of the EC by making co-decision with the Council the legislative procedure for adopting legislative acts. Corbett has characterised its impacts as follows: 'The momentum it generated was sufficient to stimulate those governments with a generally favourable attitude to European integration to take initiatives in favour of institutional reform [which] featured in the dynamic that led to the decision to convene an IGC.' (1998:217).
(22) Theoretically a legislative act could be delayed indefinitely, but the ECJ made clear in 1995 that the Council under an urgency request was entitled to adopt a regulation after the EP had withheld its opinion (Corbett, 1998:119).
(23) Since it was only in the 1987 Rules of procedure that the referral back to the relevant committee was made automatic, the EP did in fact utilise it 'rather infrequently in the early 1980's' (Corbett, 1998:120).
(24) For a detailed discussion of the bureaucratic discretion which includes the ECJ and which in some respects differs from the one employed here, see (Tsebelis & Garrett, forthcoming a; b; c).
(25) See also Golub (1997), arguing that the Luxembourg Compromise did not result in legislative stagnation.
26) Concerned mainly with the internal market (article 100A), as do the co-decision I and II procedures. The descriptive part is based on (European Parliament, 1988b).
(27) Assuming that the EP and Commission have identical integrationist preferences and 'complete info (…) the EP would make no amendments' (Tsebelis & Garrett, forthcoming, c:18). However, as a challenge to the assumption of complete information, the EP has made 4572 amendments in the period 1987 to 1993 of which the Commission has accepted 65 % and the Council 48 % (Tsebelis & Garrett, 1997:87).
(28) The descriptive is based on (European Parliament, 1994f, chapter I).
(29) The modus vivendi of 20 December 1994 signed between the EP, Commission and Council put a preliminary end to disputes concerning comitology and the EP's lack of involvement (European Parliament, 1995c:1).
(30) The EP voted on the ONP directive in the first session after the June 1994 European elections. This naturally reinforced the MEPs collective aim of demonstrating the influence of the EP and - arguably - helped it to reach the absolute majority of the component members (then 260 of 518) for a rejection of the Council's confirmed common position.
(31) The descriptive part is based on (European Parliament, 1998a, chapter B).