Available online: 24 May 2011

Transcription

1 This article was downloaded by: [University of Southern Denmark] On: 20 September 2011, At: 11:15 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Journal of European Public Policy Publication details, including instructions for authors and subscription information: Principal agent theory and the Open Method of Co-ordination: the case of the European Employment Strategy Caroline de la Porte Available online: 24 May 2011 To cite this article: Caroline de la Porte (2011): Principal agent theory and the Open Method of Co-ordination: the case of the European Employment Strategy, Journal of European Public Policy, 18:4, To link to this article: PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan, sublicensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

2 Journal of European Public Policy 18:4 June 2011: Principal agent theory and the Open Method of Co-ordination: the case of the European Employment Strategy Caroline de la Porte Downloaded by [University of Southern Denmark] at 11:15 20 September 2011 ABSTRACT This paper adapts and then uses principal agent (PA) theory to conceptualize and thereafter to analyse the EU-level development of the Open Method of Co-ordination (OMC), a crucial component of the Lisbon Strategy as a governance architecture. The PA model theorizes the continuous interaction and power struggle between the Commission ( agent ) and the member states ( principal ) in the emergence and institutionalization of the OMC. It is innovative for several reasons: first, it acknowledges that the member states and Commission interact in a PA logic prior to a contractual agreement; second, it recognizes that the principal does not only control and monitor the agent, but also (re-)defines features of the OMC via political initiatives; third, it underlines the importance of the ideational action of the agent. On the basis of the model, two hypotheses are formulated: first, that the agent will be more influential in defining the OMC when it is nascent; and second, that the principal will be more influential in the reconfiguration of the OMC. The model is tested and the hypotheses are confirmed via a longitudinal analysis (1992 to 2005) of the OMC in employment policy. KEY WORDS European Commission; European Employment Strategy; member states; Open Method of Co-ordination; policy entrepreneurs; principal agent theory INTRODUCTION Policy-specific Open Methods of Co-ordination (OMC) are crucial features of the Lisbon Strategy (LS) as a governance architecture (Borrás and Radaelli, 2011). Yet, there are still many unanswered questions about their institutionalization. Why, when and how do key actors play a role in the institutionalization of issue-specific OMCs? Which actor(s) have the most influence in defining the OMC and how does this change over time? Which theoretical framework is the most appropriate to capture the entirety of this process? The theoretical interest in the institutional genesis of the OMC, i.e., how to explain the entire process of its formation, longitudinally, has received little attention in the literature. However, there is emerging research in this area (de la Porte 2008; de Ruiter 2010; Gornitzka 2007) to which this paper aims to contribute, with a case study on the European Employment Strategy (EES). Journal of European Public Policy ISSN print; online # 2011 Taylor & Francis DOI: /

3 486 Journal of European Public Policy The ideas behind the EES have been traced back to a transnational movement, which was rooted in the Swedish labour market model (Johansson 1999). However, this does not explain its institutional genesis at European Union (EU) level, for which the advocacy coalitions approach (Schäfer 2004; van Riel and van den Meer 2002) hypothesizes that political forces across EU institutions the Commission, the European Council (EC) and the European Parliament are decisive for OMC emergence. Specifically, this approach has identified an inter-institutional social-democratic and socialist alliance as the explanatory factor for the crystallization of the EES and OMC. At best, this explains the visible tip of the iceberg at one moment in time, which is the adoption of the European Employment Title in the Amsterdam Treaty for the EES. But much more needs to be explained about how the EES was formed and how it continues to take shape once institutionalized. The neo-functionalist literature explains the role of the Commission in the development of the EES as functional and political spillover from the Economic and Monetary Union (EMU). Indeed, the policy co-ordination model of the EES was suggested by the Commission, mimicking the process for co-ordination for EMU (Hodson and Maher 2001). The advocacy coalitions and neo-functionalist approaches contribute to explaining the emergence of the EES, but they are not combined theoretically to explain the emergence and institutional evolution of the EES. The alterations that were made to the EES in the period, once it was well institutionalized, are explained as an intergovernmental initiative (Armstrong et al. 2008). However, these explanations are not combined theoretically or longitudinally to explain the institutional evolution of the EES. The literature on the EES emphasizing advocacy coalitions expects wide (social-democratic) actor involvement in its policy development. However, exploratory interviews and examination of policy documents of labour representatives and European party formations (ETUC et al. 1997; EU interview 8; EU interview 9; Party of European Socialists and European Parliament 2000) showed that the role of the European Parliament, the Party of European Socialists and the European Trade Union Confederation, with which the crystallization of the EES has been associated, was marginal (de la Porte 2008). The data clearly suggested that the EES was driven forward by the Commission and the member states (member states). When and how the Commission and the member states interacted in the formation of the EES can be theorized and then analysed via principal agent theory. Since the Commission s role in the OMC is (through an informal right of initiative) to propose common European objectives and to ensure that member states strive to comply with EU objectives in regular reports (delegated task), it is useful, conceptually, to identify it as an agent. The member states, as the principal, set the overall policy objectives of any given OMC at the yearly spring EC, and also decide on the institutional form of that OMC contract between the Commission and member states, including which tasks to delegate, and with which oversight mechanisms. Principal agent (PA) theory is adapted to the analysis of the OMC and then tested on one case: the EES.

4 C. de la Porte: The case of the European Employment Strategy 487 The research in this contribution is interpretative, since it involves putting together the evidence about a phenomenon (OMC) through process-tracing (Checkel 2005) in a way that is sensitive to chronological evolution. But it is also deductive, since a causal analytical research design is developed through PA theory. The primary data, which is triangulated, comprises official documentation from the Commission and the EC, nine in-depth semi-structured expert interviews (one to four hours) with key actors involved in the genesis of the EES, and informal contact with key member state and Commission informants in the period. The remainder of this contribution is organized as follows. First, the concepts for analysing OMC institutionalization and then the main features of the PA model, adapted to OMC analysis, are presented. Second, the model is tested on the EES, showing that the Commission was a powerful norm entrepreneur during its crystallization, until 1999, after which the member states not only increasingly controlled the Commission in the EES but entirely re-defined the EES through its own political initiatives. Finally, conclusions are drawn about the theoretical implication of the findings for the longitudinal evolution of other OMCs and of the LS as a governance architecture. CONCEPTS, THEORETICAL PA MODEL AND THE OMC In the following, some concepts will be developed to accurately trace OMC institutionalization, since it changes over time and is inherently ambiguous. Then the rationale for adopting a PA approach will be explained, following an adaptation of the PA model for the analysis of the OMC. To disentangle the features of the OMC analytically, a first conceptual distinction is drawn between OMC policy objectives and institutional model (de la Porte 2008), referring respectively to ideational components and organizational components of the LS as a governance architecture (Borrás and Radaelli 2011). Policy objectives refer to the specific policy goals or courses of action (Heclo 1974) that have been agreed via the OMC in a particular policy area. The institutional model is defined as the formal and informal rules and norms that govern the OMC policy cycle (i.e., frequency of guidelines, level of stringency of the OMC, integration of the OMC with other EU co-ordination processes, national reporting requirements, monitoring and surveillance process). The second concept introduced to trace OMC development is a temporally located stage, which is defined as a significant empirical development in the formation of a political phenomenon, in this case the OMC, which is the dependent variable (de la Porte 2008). This is crucial because the objectives and institutional model of the OMC may change over time, and because the role of the actors the Commission and the member states may differ throughout its institutionalization. While some stages of the OMC institutional development may be legally endorsed in the Treaty, others may be endorsed politically, at meetings of the EC. To analyse the OMC at each

5 488 Journal of European Public Policy significant stage is important to understand how key actors are involved in its development. The PA model is a mid-range but precise theory, which has been used for analysing processes of delegation of power to executive and judicial agents to reduce transaction costs of policy-making (Epstein and O Halloran 1999; Pollack 2002). For the EU, there has been extensive work on the analysis of conditions under which the European Commission (executive agent) and the European Court of Justice (ECJ) (judicial agent) can exercise autonomy beyond the intended contractual agreement (Bauer 2002; Pollack 2003; Tallberg 2002), and how the principal(s) respond to that. It is striking that in such cases the principals will always react by decreasing the power of the agents. Although PA analyses in EU studies have not been without critique (Kassim and Menon 2003), the value of the theoretical approach is to systematically focus on the actions of the key actors in the implementation and interpretation of a contractual arrangement. With a focus on (re-)interpretations of an agreement, it is particularly well-suited for the OMC and the LS as a governance architecture, which by definition are iteratively changeable, both their policy objectives and their institutional form. The model is particularly interested in how agents, with their preferences, can interpret a contractual arrangement in a way that increases their own power. It focuses on the activity of the agent (and principal reactions) once a contract has been agreed. Here, several adaptations will be made to the model. First, although the PA model has not been applied to study the emergence of a policy, it is pertinent for analysing the process of EU policy-making in the governance architecture of the LS, since the Commission and member states interact in a PA logic, not only during the implementation and revision of a policy, but also when it is being initiated and shaped. The use of the PA model in this way sheds light on the theoretical question of the conditions under which (and when in a policy cycle) the agent can initiate and shape policy. In this way, the PA model moves beyond the focus on how an agent interprets a contract. Indeed, even in the absence of a legal right of initiative, the Commission has some autonomy: it tries to gain power incrementally. In high-level EU policy debates, formally taking place at the Summits of the European Council, the Commission has the informal right to propose policy priorities for the member states (in Green Papers, White Papers and Commission communications). In accordance with the PA model, the Commission does this in order to gain some influence (and to increase its own power) in a particular policy area. Following Commission initiatives, member states can agree to delegate some tasks to the agent (Commission), such as analysing their education or employment policies on the basis of comparative national and/ or European statistics. The latter task can be interpreted as providing the agent with an informational advantage over the principal at one moment in time. But if such tasks are crystallized into an EU initiative EU programme, committee, statistical infrastructure or policy co-ordination cycle by the principal, then the agent s capacity to act would be institutionalized in a

6 C. de la Porte: The case of the European Employment Strategy 489 particular way, with more or less rigorous oversight procedures. A funding programme could enable the Commission to administrate projects, to decide upon fund allocation for projects and to monitor implementation. A policy co-ordination cycle would enable the Commission to produce comparative country information (according to key objectives agreed by member states) and then to analyse compliance with objectives (and in some cases issue policy recommendations) to each individual member state. Crucially, even during the emergence of a political mechanism, when the final outcome and form is not clearly defined, the member states and the Commission interact in a PA logic where the commission tries to increase its leverage, while member states resist intervention in areas of national sovereighty. The OMC is one possible outcome of a process of Commission and member state interaction. Once institutionalized, it can be characterized as a contractual relationship between the Commission and the member states, which is subject to revision, on the basis of lessons learned and the identification of new priorities (Tallberg 2002). OMCs are sector-specific, implying activity by some units within the Commission and some Council formations. For the analysis of the EES, the DG of Employment and Social Affairs (DG EMPL) is the main unit active in the Commission, while it is the Employment and Social Affairs Council (ESAC) for the member states. The PA literature recognizes that there may be multiple agents or multiple principals with diverging interests. However, key aspects of OMC development are decided on the basis of the dominant interests of principal(s) and agent(s). Even if there is divergence of positions in the run-up to a decision about the OMC, the agent(s) and principal(s) act as unitary actors during the decision-making moment (Pollack 2003). The Commission can take action through Commission Communications, White Papers, documents from institutionalized working groups and the latter through conclusions of the EC, Treaties, national position papers and documents from institutionalized or ad hoc intergovernmental working groups. OMC development also involves policy entrepreneurs, which develop long-term political ideas and policy brokers that attempt to obtain consensus on political aims (Sabatier and Jenkins-Smith 1993). Within the PA framework, all key actions related to OMC development, whether by individual entrepreneurs or policy brokers, institutionally represent either the agent or the principal. To analyse OMC development at EU level as discussed above, several types of actions by the agent and the principal are identified from the extensive PA literature applied to EU integration issues. First,the agent can develop credible expertise in areas where the principal has incomplete information; second, it can put issues on the EU agenda; third, it can lobby for its (sectoral) interest; fourth, it can shirk the principal, which is to take action that breaches its delegated tasks, and which has not been approved by the principal (Pollack 2003: 21). In addition to these actions, derived from the assumption of rational actor behaviour of the PA model, the agent can also take ideational actions, that is to develop a policy discourse around a policy problem for which a solution may be proposed (Rein and Schön 1993). In the OMC, the Commission develops ideas in policy

7 490 Journal of European Public Policy discourses (referred to as discourse framing by Bauer [2002]) in Commission communications, White Papers, documents from institutionalized working groups. Regarding the principal, the literature mainly focuses on its control and monitoring mechanisms (Epstein and Halloran 1999; Pollack 2003). In analyses of PA theory adapted to European integration, these analyses mostly focus on the interaction between a supranational agent and member states once a contract has been established. The contract defines the objective agreed by principal(s) and specifies the tasks that have been delegated to the agent to achieve that objective (Pollack 2003). If the agent shirks the principal, then the principal is expected to sanction the agent by reducing its delegated responsibilities. However, aside from control and sanctioning, the PA model does not conceptualize when, how and to what extent the principal will develop new initiatives within an existing PA process. As highlighted by Bauer (2002), there is an implicit theoretical tilt towards controllability in the PA model as applied to the analysis of European integration (p. 384). Therefore, this version of the PA model aims to capture that process: policy setting is defined as the intentional action by the member states to define policy objectives for the OMC. This concept captures the action by the member states to bargain common policy preferences for co-ordination, rather than merely to control the action of the agent. While it may draw upon ideas, it is not ideational action in the sense of the Commission s discourse framing (Bauer 2002), because it is closely tied to political preferences of the member states. On the basis of this model, two hypotheses are put forward in line with expectations of public policy literature: H1 The Commission is likely to influence the OMC more than member states when the process is in its infancy; H2 The member states are likely to influence the OMC more than the Commission when it is institutionalized. These hypotheses contribute to answering the first research question addressed in the leading paper of this collection: how did the LS (ideational and organizational components) emerge? Once created, how was it maintained and adapted over time (Borras and Radaelli 2011)? In the following, the hypotheses will be tested by analysing the longitudinal development of the EES from 1992 to In the operationalization of this model, successful action is considered to be attained when an actor ( agent or principal ) successfully accomplishes its intended aims during OMC formation, by using the types of action at its disposition. ANALYSIS OF EES INSTITUTIONALIZATION VIA THE PA MODEL Six key stages have been identified in the emergence and development of the EES, from 1992 to Table 1 summarizes the empirical outcome of the OMC. It includes a specification of when the stages were institutionalized

8 Table 1 The institutionalization of the EES Stage Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 6 Time June 1997 November Aspect Concrete empirical content Decisive actor in change Policy objectives and institutional model Political objectives and loosely defined administrative tasks Institutional model Policy objectives Institutional model Policy objectives Policy objectives Formal (legal) establishment of the Employment Policy Process First set of employment guidelines Use of recommendations Overarching intitative objectives in the EES; creation OMC Agent Principal Agent Agent Principal Principal Main action Discourse framing Contract definition Discourse framing Shirking Setting policy direction Poltical party leadership in Council and Commission leadership Majority of conservative prime ministers (PM) in Council; strong (Delors) Commission Majority of social democratic PM in the Council; weak (Santer) Commission Majority of social democratic PM in the Council; weak (Santer) Commission Majority of social democratic PM in the Council; weak (Santer) Commission on brink of resignation Majority of social democratic PM in the Council; weak (Prodi) Commission Change to the policy objectives and institutional model of the EES Setting policy direction, redefinition of EES contract Majority of conservative PM in the council; weak (Barosso) Commission Source: Authors interpretation; Tallberg and Johansson (2008) for political leadership in Council; Kassim and Menon (2010) for strength of Commission leadership.

9 492 Journal of European Public Policy (time), and a description of the concrete empirical outcome (aspect). Second, it points to the main findings on the basis of the PA analysis, indicating when it was the agent and when it was the principal that was the main actor (decisive actor in change) in an EES stage. Political party configuration among the member state representatives in the Council and the strength of Commission leadership are noted, since they may contribute (as intervening variables) to explain shifts between agent and principal influence on the OMC over time, as well as shifts in the policy direction of the OMC. Downloaded by [University of Southern Denmark] at 11:15 20 September 2011 Stage 1: Discourse framing driven by the Commission as a norm entrepreneur In the aftermath of the 1991 Maastricht Treaty that defined the EMU, a severe economic crisis hit Europe: economic growth was low, the unemployment rate was high (10.4 per cent) and the employment rate was low (60.8 per cent) (Eurostat 2006). The member states reflected their concern about this at the Edinburgh Council, in 1992 (European Council 1992). This was the springboard used by Jacques Delors perhaps the most powerful President of the European Commission ever to obtain a mandate to prepare a policy paper (Kassim and Menon 2010; Ross 1995). In a short policy paper, Delors proposed a new idea: to increase employment rates, supported by active labour market policies (ALMP) as a response to low growth. The novelty in the policy discourse developed by Delors was to shift attention away from unemployment rates, dominant in member states political and public debate, towards employment rates (Delors 1993). This employment policy discourse was endorsed at the following European Summit by member states, where the Commission was given a mandate to develop a strategy for boosting growth and employment in the EU (European Council 1993a). The Commission worked on the White Paper for Growth, Competitiveness and Employment, building on the main policy solution in Delors s initial policy paper, which was to shift focus from decreasing unemployment to increasing employment (Delors 1993). It proposed a quantitative benchmark to increase the employment rate of the EU to 70 per cent and a loose policy co-ordination (institutional model) as the means to implement this strategy. After intense lobbying by Delors and his team (EU interview 3; EU interview 4; Ross 1995), the member states agreed on the policy objectives for employment policy in the White Paper: to invest in vocational training; to develop more flexible work organization patterns and a wage policy that encourages job creation; to reduce non-wage labour costs; to improve the effectiveness of labour market policy by shifting from passive to active labour market policy; and to develop measures for vulnerable groups (European Council 1993b). At the Essen European Council (European Council 1994), the last Summit attended by Delors, these objectives and loose co-ordination were accepted by the member states. While there were diverging views within the Council ( multiple principals ) in the run-up to the European Council, at the end even those opposing binding provisions in employment policy, in particular

10 C. de la Porte: The case of the European Employment Strategy 493 Helmut Kohl (Christian-Democratic Union, Germany) accepted the objectives and the ad hoc co-ordination process because it was loose (EU interview 4). In fine, the Council acts in unison, despite diverging positions. The definition of the employment policy discourse and the accompanying institutional model had been driven forward by the Commission as the agent, on the basis of Delors s ideas that were proposed in 1993 and then developed in the Commission White Paper. The agent used ideational action and informal agenda-setting power to successfully define the core direction and the aims of EU employment policy. This was possible because of unusual circumstances: the (exogenous) economic crisis and the (endogenous) strong President to the European Commission. Stage 2: Contract definition (Employment Title) by the principal In January 1995, Jacques Santer, a Christian Democrat, replaced Jacques Delors as President of the Commission. He was intentionally nominated by member states because he was a weak leader (Kassim and Menon 2010: 20). The Santer Commission pursued actions on the basis of the legacy left by Delors (Kassim 2010). As explained in an interview: The Commission supported the development of an employment process behind the scenes since the mid-1990s, well before the Amsterdam Summit in The main active supporting units within the Commission were the DG EMPL and the DG General Secretariat. It built on the legacy of the 1993 Delors White Paper. The Commission represented a stable pillar of support to the development of the EES. (EU interview 4) Thus, the agent actively supported the development of a binding process and objectives before it was institutionalized in the Treaty. At the same time, Allan Larsson, previously head of the Swedish socialist party, became director-general of the DG EMPL (1995 to 1999). Larsson has been identified as a key entrepreneur in the ideational and policy development of the EES (Barbier 2004; EU interview 1; EU interview 2; EU interview 3; EU interview 4; EU interview 5; EU interview 6). The DG EMPL, under the leadership of Larsson, proposed the development of an employment and labour market committee (EMCO), subordinate to the ESAC, and akin to the Economic and Financial Committee subordinate to the Economic and Financial Affairs Council. EMCO would become a key feature of the EES institutional model and its creation was not per se politically sensitive, particularly as it institutionalized member state (civil servant) participation in EU employment policy. The decision to establish EMCO was adopted in 1997, as it ensured intergovernmental steering in the development of the EES. The decision to legally institutionalize the Employment Title in the Amsterdam Treaty was a more sensitive issue, however. In the negotiation of the Amsterdam Treaty, many member states, particularly those represented by social democratic prime ministers, aimed to obtain a legal basis for

11 494 Journal of European Public Policy employment policy. Particularly for the new member countries that entered the EU in 1995 Sweden, Finland and Austria employment policy co-ordination at the EU level was presented to the national electorates as a means of safeguarding social standards (EU interview 2; EU interview 4). In Sweden, Allan Larsson worked, with Prime Minister Ingvar Carlsson (1994 to 1996) in this direction, and presented a paper, A vision for the IGC (Intergovernmental Conference). In the paper, the idea of developing an employment Union alongside EMU was presented: the creation of an Employment Union would be the expression of a common European commitment to give a central role to the battle against unemployment (Larsson 1995: 5 6). Larsson had lobbied actively in DG EMPL, but also vis-à-vis the Council (representing Sweden) to legitimize the EU for Sweden. This finding supports recent research, which shows that Commissioners Director Generals could be expected to be driven by similar priorities are driven by their sectoral portfolios as well as national agendas (Wonka 2008). In this case, Allan Larsson primarily represented the Commission, but his policy entrepreneurship was also aimed at providing a positive message to the Swedish electorate about the EU. In other countries with left-oriented governing coalitions, governments supported the development of a legal status in the Treaty for the EES, but it was possible to obtain an agreement among the EU-15 because of shifts towards left-wing governments among the big EU countries, in particular Blair (Labour Party, UK) and Jospin (Parti Socialiste, France) (Goetschy 1999; Lightfoot 2003; Schäfer 2004; Tallberg and Johansson 2008). Kohl (Christian Democratic Union, Germany), who had been opposed to the Employment Title since Essen, and Aznar, representing the conservative coalition in Spain, agreed to include the Title in the Treaty owing to its soft law character. It is notable that Wim Kok, then Prime Minister of The Netherlands and later to be a key entrepreneur in setting the policy direction of the EES and the LS in 2003, signed the Treaty of Amsterdam for his country. The Amsterdam Treaty (European Communities 1997) was endorsed at the Amsterdam EC, which concluded the IGC (European Council 1997a). The Title represents a legal commitment to the political objective of a high level of employment through an iterative policy cycle. However, in the agreed Employment Title, only participation in the co-ordination cycle was rendered mandatory; total compliance with the policy objectives was encouraged, but could not be sanctioned. While the objectives of EU employment policy co-ordination built on ideas developed by the agent, it was the action of the principal (agreement on a contractual arrangement ) that was decisive at this stage. Political circumstances, as an intervening variable, facilitated this but were not decisive. Stage 3: Entrepreneurial action of the agent and employment policy objectives During stage 3, the Commission worked for the development of employment policy objectives, on the basis of ideas in the White Paper (EU interview 2).

12 C. de la Porte: The case of the European Employment Strategy 495 The work of DG EMPL under the leadership of Allan Larsson was presented in a Commission communication that was organized around four themes, in the following order: entrepreneurship; adaptability; employability; and equal opportunities (European Commission 1997). In addition to policy objectives developed to increase the employment rate, the Commission communication presented contained the same quantitative employment rate target 70 per cent for the EU (European Commission 1997) that had been included in Delors s 1993 White Paper (European Commission 1993). The dominant thrust of the Commission s strategy, defined by DG EMPL but also supported by the DG General Secretariat and the DG ECFIN, was the idea to place more pressure on member states to increase their employment rates. The employment rate across the EU at this time was 60 per cent and the unemployment rate was 10 per cent (Eurostat 2006). In the Council, all countries were eager to take measures to enhance employment rates expected to support economic growth (EU interview 2). Nevertheless, because employment policy is an area of national sovereignty, the Commission needed to work together with the member states (lobbying) to obtain support for the policy objectives it had developed. The Commission President, Santer, worked with Jean-Claude Juncker, then acting President of the EC, for the preparation of an extraordinary meeting of the EC ( Jobs Summit ) to agree on the EEG. Both were from Luxembourg and both emanated from a Christian-Democratic tradition, from which they derived the same frame of reference and policy perspective in employment policy. While Santer collaborated with Juncker for the practical preparation of the Summit, it was Juncker, representing the principal, who engaged actively in policy brokering in the run-up to the Jobs Summit (November 1997). An interviewee pointed out that Juncker in particular played a crucial role in achieving agreement on the content of the first set of guidelines. He toured all capital cities to get a sense of which political issues were consensual (EU interview 3). After this, Juncker changed the order of the policy objectives proposed by the Commission (European Commission 1997) in line with member state political priorities, whereby employability was placed first, entrepreneurship second, adaptability third and equal opportunities fourth (EU interview 2). All guidelines proposed by the Commission were accepted, but the 70 per cent employment rate benchmark were not endorsed by the Council. The objectives for the EES in the order proposed by Juncker were endorsed at the special summit initiated by Luxembourg at the end of 1997 (European Council 1997b). The guidelines were developed by the European Commission, but would not have been accepted by member states had it not been for the role of Juncker as policy broker. In the period there was a low level of tension between the principal and the agent that were working towards common objectives that were in the interest of both actors. As Prime Minister of Luxembourg, Juncker represented the EC, but he also endeavoured to pursue labour market reform in Luxembourg. Like Larsson, Juncker had multiple playing fields, but his

13 496 Journal of European Public Policy action was for the principal in this case. After the endorsement of the guidelines, Juncker used the EES to introduce reforms in Luxembourg, notably incentives for the development of female labour market participation and childcare institutions (EU Interview 2). The findings of the initial crystallization of the EES in stages 1, 2 and 3 are in line with hypothesis 1: the Commission s scope of action was considerable during the initial genesis of the EES. Indeed, all the work both on the form and the content of the EES had been carried out by the European Commission prior to the member state adoption of the Treaty, via discourse development and, crucially, by lobbying towards high level member state representatives. The social-democratic majority in the Council, which has been highlighted in the literature as the main reason for the adopted of the European Employment Title, is only an enabling condition for its adoption. This full definition of the EES as a formal contract in the Amsterdam Treaty, was decided by member states. Stage 4: Recommendations in the EES: from agent shirking to principal control During stage 4, the Commission was keen on gaining power within the EES, which it interpreted in the view of obtaining more power (vis-à-vis member states). The Employment Title that defines the EES co-ordination cycle stipulates that recommendations may be used if deemed necessary. The recommendations represent an instrument with which the Commission can pinpoint acute policy problems (or misfit with EU objectives) in individual member states. However, the recommendations must be approved by the ESAC prior to being issued to individual member states, a mechanism to avoid too much Commission intervention in domestic policies. During the first two rounds of implementation of the policy cycle of the EES ( and ), organized and monitored by the Commission (ensuring national reporting requirements by member states would be met, assessing individual country progress, organizing peer review), individual recommendations had not been issued to any of the member states. During stage four, particular (exogenous) circumstances instigated the Commission to launch recommendations to member states within the EES. In 1999, the whole Commission was to step down for mismanagement, following an audit organized by the European Parliament. Prior to his resignation, Commissioner Flynn of the DG EMPL sought to strengthen the EES policy co-ordination cycle, by mimicking the practice in the Broad Economic Policy Guidelines (BEPG) of making country recommendations during each full policy cycle (EU interview 3). Flynn marked the EES by making individual country recommendations to all member states in September 1999, without authority to do so from the member states. This is a clear case of shirking by the agent, since it was against the procedures that had been established for Commission action within the EES. However, since the Commission was to resign, it was not concerned with the consequences of its action. Following the shirking, member states, by individually negotiating

14 C. de la Porte: The case of the European Employment Strategy 497 with the Commission, softened the recommendations issued to their respective countries (Barbier 2004; EU interview 5). After this, the ESAC accepted this first round of recommendations as a Council decision (EU interview 3; EU interview 5). In the aftermath of this event, the use of the recommendation tool was institutionalized as a regular feature of the EES, where the Commission was obliged to accept the oversight mechanisms established by the ESAC for the use of the recommendations in the EES. The recommendations were to be based on wellfounded analyses and could not be polemical or contrary to what the member states identified as key problems (Barbier 2004). During this stage, the Commission interpreted the (EES) contract in its favour, while the member states responded by curbing its power and introducing oversight mechanisms, which is in line with the predictions of PA literature. Stage 5: Council entrepreneurship renders the EES more intergovernmental Following the political crisis of Commission resignation in 1999, Romano Prodi was selected as President of the Commission and comprehensive administrative reform was undertaken, rendering the Commission weaker (Bauer, 2008; Kassim 2008; Peterson 2008). Under these circumstances, the Council during the Portuguese Presidency (the first half of 2000) aimed to strengthen the EES objectives and institutional model as part of a global socio-economic strategy. António Guterres, Prime Minister of Portugal, was the main policy entrepreneur behind the LS, and Maria Rodrigues, his advisor, was an important policy broker (EU interview 7; Rodrigues 2002). In 2000, the annual economic growth rate was 3.9 per cent of GDP in 2000, the highest it had been since the beginning of the recession in 1991 (Buechs 2007; de la Porte 2007) and the employment rate was 63.4 per cent. The mood among member state representatives in the (predominantly socialdemocratic) Council was therefore optimistic. In the preparation of the Lisbon Summit, the Portuguese Presidency disseminated, in January 2000, a strategic document to member states, European institutions and interest organizations, to test which ideas were consensual. The document drew up a strategy for the EU in competitiveness, employment and social policy and proposed that policy co-ordination (not yet labelled the OMC) should be used in a number of policy areas to achieve the new aims. In the area of employment policy, the document called for job creation, especially in the services sector, the inversion of the trend towards early retirement, an increase in female labour market participation, reconciliation of work and family life and life-long learning. Finally, it proposed that quantitative employment rate targets 70 per cent overall employment rate and 60 per cent female employment rate should be adopted in the EES to sustain economic growth and social cohesion (Council of the European Union 2000). The principal was using the ideas behind the EES, which were initially set out in the Delors White Paper and

15 498 Journal of European Public Policy further developed in the EEG by the European Commission in In February, the Commission circulated a document, highlighting the same policy challenges and responses as the Portuguese Presidency in January, but without any novel input (European Commission 2000). There was full support among member states for the aims proposed by the Portuguese Presidency: to stimulate economic growth; to create more and better jobs; and to strengthen social cohesion in a global economy (European Council 2000). Second, the overarching employment rate benchmarks 70 per cent overall employment rate and 60 per cent female employment rate proposed by the Council were adopted as a core feature of the EES. When the Commission had proposed the employment rate benchmarks in the Delors White Paper (stage 1) and later in the draft employment guidelines (stage 3), they had not been adopted. Third, the OMC was launched as a policy co-ordination instrument, to be applied to different policy areas (de la Porte 2002). Fourth, an annual spring summit was to review the yearly socioeconomic developments across the EU and to adapt its socio-economic aims. The decision-making during this stage was driven by the member states through policy- and agenda-setting, as well as brokering among member states, while the Commission was on the backseat. Stage 6: Change to the EES: a voluntarily initiated member state triumph In 2002, the EES was revised on the basis of demands from civil servants within EMCO: they wanted a less bureaucratic and more result-oriented EES, which should be aligned more closely to the economic policy co-ordination cycle and objectives. It was this agenda that entirely fuelled the Commission communication of July 2002, written by the DG EMPL. During this stage, the Commission ( agent ) was responsive to the priorities of member states, with little room for including its own ideas. The Communication comprised 11 objectives organized around three core aims: to achieve full employment; to raise quality and productivity at work; and to promote cohesive and inclusive labour markets (European Commission 2002). Compared to stages 1 and 2 of EES crystallization, when the Commission used ambiguity in its favour and was successful as a norm entrepreneur, and stage 4, when it shirked the member states, the policy contribution of the Commission at this stage was minor as the altered policy objectives simply reflected the wishes of member states, whose interests were systematically institutionalized in EMCO (for which the Commission acted as secretariat). This signifies that the more the EES is institutionalized, the more the member states define its political agenda, although it may be proposed by the Commission. Shortly after this, some member states expressed concern about the EU s economic growth rate, which had fallen from 3.9 per cent of GDP in 2000 to 1.1 per cent in 2004, and its employment rate, which was virtually unchanged, at 64.3 per cent (Eurostat 2006). Wim Kok, Prime Minister of

16 C. de la Porte: The case of the European Employment Strategy 499 the Netherlands, initiated a working group European Employment Taskforce (EET) voluntarily and outside of the EES process to re-define its policy objectives and institutional model. The EET developed policy aims that were set out in a report (Kok 2003), which were endorsed by the EC. The novel objectives of the Kok group which were adopted by the EC focus on investment in human capital, the development of skills and the revision of education systems, whereas the objectives proposed by the Commission had focused more on policies for female labour market participation and the reconciliation of work and family life. While the policy objectives to some extent reflect the change in political coalitions in the Council (conservative liberal), such an interpretation ignores the fact that the main initiator of the EET, Wim Kok, was a social-democrat. Furthermore, party priorities across the left right political divide were converging towards a strategy focusing on increasing employment rates and ensuring continuous human capital development (Buechs 2007). The redefinition of the EES took place in two phases: first, the member states defined the EU employment policy agenda (in EMCO); second, political actors at a higher level (the Kok group), ignoring the work of the civil servants in EMCO, redefined the policy objectives and the institutional model of the EES through a political initiative. After 2003, procedural ambiguity (Borrás 2009) of the EES was exploited by the member states to redefine the EES. The ongoing reform within the Commission during this period (Bauer 2008; Kassim 2008;) reduced the power of the Commission power in several areas, including employment policy. Ultimately, the EES as a whole became more intergovernmental after 2000, and this trend was strengthened after These findings confirm hypothesis 2, according to which the principal is expected to become more involved in the OMC after its institutionalization, because only then does it represent a genuine political opportunity structure for member states. While the principal had reduced the power of the agent in stage 4, since it clearly used the EES to its advantage, the dynamic at hand in stages 5 and 6 is completely different. It is driven politically by high-level actors, with no genuine ideational input of the Commission. These findings suggest that the policy-setting action of the principal should be taken seriously in PA modelling, particularly in the political setting of the LS as a governance architecture. CONCLUSION This contribution has incorporated ideational, political and longitudinal elements into PA theory, in order for it to be applicable not only to clearly defined legal contracts, but also to nascent and evolving political contracts, such as OMCs and the LS as a governance architecture. The model has been tested on the emergence and evolution of the EES. The findings are that the Commission was more influential than the member states on the EES during its initial genesis because of the economic crisis which affected Europe and

17 500 Journal of European Public Policy because the EES did not represent a threat to member states. When the Commission, without consulting the member states, issued individual country recommendations to member states in 1999, member states reacted to Commission shirking by implementing control mechanisms for use of recommendations. This is not surprising: the PA model predicts that the principal reduces the power of the agent, if it takes steps which are beyond the delegated tasks. However, the development of the EES after this period, driven by the principal, takes place through a political rather than a control logic. The principal, through a voluntary political initiative, redefined the EES entirely after 2003, when the Commission was weak. From a theoretical perspective, the inclusion of political activity policy-setting of the principal makes PA theory relevant for analysing the decision-making and institutionalization of the LS as a governance architecture, as well as issuespecific OMCs the longitudinal perspective, beginning during the formation of a political phenomenon, takes account of the ideational action of the agent, and has crucial explanatory power for the genesis of political instruments in the EU. Biographical note: Caroline de la Porte is associate professor at the Centre for Welfare State Research, Department of Political Science and Public Management, University of Southern Denmark. Address for correspondence: Caroline de la Porte, Center for Welfare State Research, University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark. cpo@sam.sdu.dk ACKNOWLEDGEMENTS A draft of this paper was presented at the CONNEX conference, Mannheim (2008) and at the ECPR conference, Potsdam (2009). Special thanks to Jean- Claude Barbier, Susana Borrás, Patrick Emmenegger, Joseph Melchior, Christilla Roederer-Rynning, the editors of this collection and two anonymous referees for incisive comments on this paper. REFERENCES Academic sources Armstrong, K., Begg, I. and Zeitlin, J. (2008) EU governance after Lisbon, Journal of Common Market Studies 46(2): Barbier, J.C. with Samba Sylla, N. (2004) La stratégie européenne pour l emploi: genèse, coordination communautaire et diversité nationale, Report for the French Labour Ministry, Paris. Bauer, M. (2002) The Commission and the poverty programmes, Journal of Common Market Studies 40: Bauer, M. (2008) Diffuse anxieties, deprived entrepreneurs: Commission reform and middle management, Journal of European Public Policy 15(5):