With the Amsterdam Treaty of 1997, employment policy became an important issue on the European level. Taking into account different institutional frameworks and economic conditions in the EU member states, the open method of coordination aims to create convergence by defining common employment objectives while leaving the policy measures in national responsibility. From the very beginning, the European Employment Strategy (EES) has focussed on young people's integration into employment, because youth unemployment is an unpleasant phenomenon in all European countries. For youth experiencing a failed integration into the labour market, perpetual disadvantages can be expected. Considering the demographic change, a large group of young but disintegrated people constitutes a potential threat for the economic growth.
Assessing the general effects of the European Employment Strategy within a comparative approach means to be confronted with a couple of analytical pitfalls and challenges. First, one needs to specify the concrete dependent variable: Does 'having an effect' mean that certain policies are created in accordance with the overarching goals of the EES (the output dimension)? Or, does the researcher refer to empirical changes originated from these policies (the outcome dimension)? Second, the question for effects always implies the proof of causal mechanisms, which again require certain methods able to investigate causality. These methods mostly rely on probabilistic theory and are only applicable having a large number of observations at one's disposal. Even if the European Union currently has 27 members, this number is much too small for this purpose. The third reason is of practical nature and refers to the huge amount of information that has to be collected in order to investigate EES effects comparatively. Finally, some of the basic assumptions of empirical research are violated when comparing European countries. These are primarily a certain degree of homogeneity and of independence of the observations.
The review of the research literature regarding the effects of the EES or the "Open Method of Coordination" (OMC) reveals that systematic comparisons between more than two member states remain exceptions (Mailand 2008). This is valid for both the analysis of the output as well as for the outcome dimension and reflects the problems described above. Furthermore, most of the studies identify significant differences in EES effects between countries. Finally, the EES seems to have more impact on national policy making processes themselves - that includes problem perception, agenda setting, and decision making - than on the content of policies (Zeitlin/Pochet 2005; Mailand 2008). All these findings are based on investigations of employment or welfare policies in general (e.g. Heidenreich/Zeitlin 2009). Regarding school-to-work transitions, no scientific assessment of the impact of the EES is available. This article aims at partly filling this gap by comparing the performance of the youth labour markets of EU member countries at the start of the Lisbon process in the year 2000 and the performance in 2008, where the latest data are available. Given the above mentioned limitations on identifying causal mechanisms, conclusions on the effects can only be obtained while using the heuristic method.
According to very different institutional settings in European countries, transitions from school to work in Europe are very diverse. This is not only shown in various levels of youth unemployment, but also in different pathways, which school leavers take entering the labour market. Transitions are influenced by certain aspects of the education and training systems - such as the degree of standardisation or stratification (Allmendinger 1989) or the vocational specificity (van der Velden/Wolbers 2003) - and of the employment system - like employment protection (van der Velden/Wolbers 2003; Breen 2005; Wolbers 2007). Apart from these institutional dimensions, one has to take into account general labour market conditions, which are still different despite the common market of the European Union. Policy activities also vary between countries, because they have to be in line with institutional peculiarities - even if labour market policies are increasingly coordinated on the European level.
Regarding policy activities related to school-to-work transitions, high and persistent youth unemployment in the European Member States caused the relative importance of youth labour market issues within the European Employment Strategy. The introduction of the Lisbon process marks a shift from a passive to pro-active policy scheme. Whereas in the 1990s the EU emphasised labour market flexibility as an instrument for tackling youth (long-term) unemployment, the EU started stressing the importance of increasing and preserving employees' human capital (Pastore 2007). In tackling long term youth unemployment, labour market flexibility turned out not to be sufficient. Therefore, the core of the EU strategy against youth unemployment included increasing labour market flexibility as well as a shift to more pro-active schemes, such as the reform of the education and training systems. The main objective of the European Employment Strategy is that "educational systems should become of a higher quality, more inclusive to reduce the dropout rate, homogenous to other EU countries to favour labour mobility, flexible to allow young people to better find their best match" (Pastore 2007: 5). The EU also follows the principle of duality, which means emphasising the importance of providing job training together with education in order to favour smoother school-to-work transitions. In a nutshell, the focus of the European Employment Strategy - originally concentrated on flexibility and activation - is increasingly accompanied by targeting matters of education systems.
The objective of this paper is to compare the policy activities and the youth labour market performance in 15 EU member states1 regarding school-to-work transitions. These transitions are seen as a result of the goals of the Lisbon process. For this purpose, the national differences in youth labour market and individual transitions between education system and labour market are examined (section 2). Then, the institutional settings of education and employment systems are described briefly (section 3), because they constitute the framework in which policy programmes take place. Finally, the targets of the European Employment Strategy are described and resulting national policies are analysed (section 4). To some extent, this approach reflects the procedure of the European Employment Strategy itself, because it is geared towards the definition of common empirical quantified goals, while at the same time the concrete definition of national policies is left to the member states.
2 School-to-Work Transitions in the European Union
In sociological research, many indicators were proposed that aimed at measuring and assessing the quality of school-to-work transitions. On an aggregate level, this means using classical indicators such as the youth unemployment rate, the employment rate of young people or average entry wages. Within the last decades, the OECD developed an extensive body of aggregate-level indicators beyond these classical measures. The transition period is often calculated by taking the duration between the school leaving age - referring to the average age at which 50% of an age cohort have finished school - and the median job entry - defined as the age at which the employment-population ratio reaches 50%. By repeated measurements within comparative report designs across approximately 30 OECD-countries, these indicators comprehensively outline an accurate empirical picture of school-to-work transitions in each country studied. On the individual level, the very basic indicator is the first transition into employment (Russell/O'Connell 2001), which is not necessarily meaningful. In order to circumvent this problem, researchers tried to detect the crucial status change by accurately constructing concepts such as the "first significant job" that lasts at least six months, for example. This limitation serves to exclude very short, probably erratic employment periods that are of secondary relevance. However, the determination of the time period that has to be regarded as 'not significant' remains to a large extent arbitrary, which cannot be avoided even by increasing quality of available data. According to the purpose of this paper - that is the examination of EES effects on national level outcomes - the aggregate indicators seem to be appropriate, because they are available for the whole time period and beyond, and because individual indicators lose their advantage when being aggregated on the national level.
For a first instance, we use a simple aggregate measure in order to assess the effect of the European Employment Strategy on the youth labour markets in EU member states. The relative youth unemployment rate is calculated by simply dividing the unemployment rate of those younger than 25 by those who are 25 years and older. The relative youth unemployment rate has the advantage that youth unemployment is expressed in relation to overall employment and, therefore, it provides a kind of control for general economic conditions. The comparison of the relative youth unemployment in 2000, when the Lisbon process was initiated, and of the rate in 2008, should provide a success indication of the objective to enhance the situation of school leavers.
Figure 1 shows that the performance of European countries regarding relative youth unemployment differs remarkably. In 2000, this indicator ranges from 1.1 in Germany - which means that youth unemployment is only slightly higher than the general unemployment rate - to 3.6 in Italy - which means that youth unemployment is...