Abstract
This study sets out to describe and explain variation in cross-level policy coherence between central level policy recommendations by the Stichting van de Arbeid (Dutch Labour Foundation) and decentral level Collective Labour Agreements (CLAs) in Dutch labour relations. Cross-level policy coherence can be understood as a problem of policy implementation.
... read more
However, Dutch labour relations are characterised by a combination of (a) multilevel governance – collective decision-making takes place on both the central and decentral level, and (b) soft policy – Labour Foundation recommendations are not legally binding. These characteristics raise two important questions: (1) to what extent can we expect decentral compliance with central level recommendations, and under what conditions? (2) To what extent are existing theories of policy implementation that assume state centric systems – which are characterised by binding political decisions and institutionally embedded systems of administrative procedures for political control – appropriate for the study of multilevel systems with ‘soft’ policies? To explain cross-level policy coherence a model-guided approach is applied. The starting point is the standard model of policy implementation. This model is adjusted and supplemented using ‘middle range theories’ on policy processes in multilevel systems. Cross-level policy coherence is defined as coherence between central level recommendations and the individual (policy positions, concessions in negotiations) and collective outcomes (negotiation agenda, CLA-arrangements) of choices of decentral level CLA-negotiators. Hypotheses are derived, predicting effects on cross-level policy coherence of (1) collective support for recommendations among peak organization administrators, (2) preference loss of individual peak organization administrators in the central level negotiations, and (3) government salience. To test the hypotheses two types of data are analysed: (1) decision-making data on (a) central level recommendations in two different social pacts, and (b) eight CLAs, and (2) longitudinal data on the presence of 75 types of arrangements about employability and work-life balance in 138 CLAs reached in the period 1996-2003. It was found that Labour Foundation recommendations have more than a symbolic effect. When CLA-negotiators hold a position that is not in accordance with a central level recommendation they are more likely to make concessions during the negotiations, than when they hold a position that is in accordance with a recommendation. In addition, the preferences and priorities of government affect decision-making processes between employers and employees, both at the central and decentral levels. The adjusted model of implementation in Dutch labour relations, which assumes that actors trade off preference loss and reputation loss, proves to be useful for future research. However, the composition of preference loss functions and reputation loss functions might need to be adjusted and supplemented. Furthermore, the effects of collective support among peak organization administrators, individual preference loss of peak organization administrators, and government salience should be tested using a larger and more consistent dataset.
show less
