Although 'integration with retention of "own" culture' has ceased to be the dominant policy principle in Dutch minority policies for quite some time now, there are still remarkably many ethnically specific policy arrangements in the Netherlands. To explain this contradiction this paper introduces an administrative mechanism: the logic of making policy categories conflicts with the logic of policy implementation. The use of 'avoidant categories' – a particular type of policy category discussed in this paper – creates an administrative opportunity structure that unintendedly promotes ethnic fragmentation instead of integration in policy implementation. We illustrate the working of this mechanism in a comparative perspective; the Netherlands is not unique in this respect and experience in other countries is instructive.

In the discussion about the introduction of a directly elected mayor in the Netherlands the issue of election procedures is largely ignored. Election procedures, however, are not neutral. Five different elections procedures that can be used to elect a mayor (simple majority, two round system, alternative vote, Cooms rule and the adjusted Copeland rule) are analyzed using seven different criteria (selecting majority winners, selecting condorcet winners, not selecting condorcet losers, monotonicity, strategy proofness, simplicity and costs). With respect to most of the criteria used, the adjusted Copeland score rule outperforms the other procedures mentioned.

The creation of quasi-autonomous organizations has spread throughout the western world. Flanders and the Netherlands both have a longstanding tradition of putting policy execution at arms' length, thereby creating so-called voi's and zbo's respectively. This raises the question whether there are comparable trajectories, forms, developments and political considerations. By comparing the developments in both countries the authors seek an answer to this question.

The establishment of quangos was very popular until the mid 1990s. However, in both countries a countermovement can be seen. The creation of quangos is believed to lead to problems for political control. Both governments have taken several measures to solve these issues.

The comparison shows that there are indeed many similarities, but also reveals interesting differences. There is no comprehensive convergence between the two neighbours. These differences are the result of differences in the politico-administrative system and the institutional culture.

Alfred Hirschman has analyzed consumers reactions of 'voice' and 'exit' with respect to the functioning of public and private organizations. He has argued that the value of 'voice' to articulate dissatisfaction with respect to public and private organisations is often misunderstood. Especially with respect to so-called complex goods, voice is far more fruitful to explore the norms for 'best practice' and to stimulate the improvement of the organisation than exit. Insurance might be seen as a complex good. From the perspective of Hirschman we analysed the exit- and voice options of citizens with respect to private and public health insurance. We argue that while private insurance offers more exit-options, the quality of the voice options in private insurance might be considered better. However, the opportunities to articulate disagreement or dissatisfaction on an individual level are few all together. Public health insurance does have formal channels for collective voice, but these do not result in real influence on relevant themes. With the reorganization of the Dutch welfare state, much has been invested, ideological and practical, in the increase of exit options. However, dissatisfaction often does develop when people have become ill and when they are dependent upon the insurance. In that situation they often are not able to walk out to a competing company or to perform voice. Against this background we argue that the organisation of collective voice in health insurance is very important.