14 The Dutch Policy Agendas Project

The Dutch Policy Agendas Project was initiated in the summer of 2006, shortly after an ECPR workshop on agenda-setting organized by Christoffer Green-Pedersen and Arco Timmermans in April 2006 in Nicosia, Cyprus. Since then, the project has been directed by Gerard Breeman and Arco Timmermans, and it moved from coding national executive agendas to the agendas of political parties, the legislature, the media, local executives, and party think tanks.

A central characteristic of the Dutch Agendas Project’s approach is its explicit combination of quantitative and qualitative methods. Besides quantitatively mapping and explaining the up-and-down patterns of the policy agendas over time, we want to put meaning to these patterns for both academics and practitioners. As we will illustrate at the end of this chapter, we analyzed, for instance, how the development of attention to security issues in the past twenty-five years revealed that the securitization agenda is the result of both proactive long-term policy investment and short-term reactive policy decisions with a crisis element in them.

The Dutch parliament (Staten-Generaal) consists of a first and a second chamber. The 150 members of the second chamber are directly elected based on proportional representation. There is just one national electoral district and there is practically no electoral threshold, which means that small parties can easily enter the second chamber. After the general elections of 2017, there were thirteen parties in the second chamber. The seventy-five members of the first chamber, the Senate, are elected indirectly. Contrary to senates in many other countries, the Dutch Senate cannot formally propose or amend legislation, but only accept or reject.

Thus far it has always been necessary to form a coalition government, which typically works with a written coalition agreement that forms the political agenda of the government, and which usually also commits the supporting parties in the second chamber. Traditionally, the three large parties are the center-left PvdA (Labor; social-democrat), the centrist CDA (Christian-democrat) and the center-right VVD (liberal-conservative). Since 2000 new parties have entered the stage, of which most notably are the party for the elderly, the animal rights party (“Partij voor de Dieren”), and the provocative anti-Muslim and anti-EU party PVV (party of freedom) led by Geert Wilders.

The traditional Dutch model of consensus (Lijphart, 1968) has come under pressure. Despite mechanisms of coalition governance such as written coalition policy agreements and arenas for political conflict management, the branches of parties in government have become more exposed to leadership battles. Traditional political parties experience function loss, and the politicization of issues such as immigration and the European Union has turned out to be a fruitful activity for populist parties. In this development, new institutional mechanisms of representation and interest representation are employed, such as referenda and negotiations with social and economic stakeholders in order to extend the basis of societal support for government policies.

In the Netherlands, institutional friction plays an important part in coalition politics, and the Dutch Policy Agendas’ datasets cover all the “stages,” of the policy cycle, from input agendas to output. Table 14.1 presents the datasets developed in the Netherlands.

In 2016, the codebook and Dutch datasets were adjusted to come into line with the international CAP coding scheme. The Dutch codebook thus contains twenty major topic categories and each of these is further subdivided into more detailed subtopics, in total 226 subtopics. All datasets are constructed with historically consistent topic categories that do not change over time. We paid attention to coding items consistently in terms of the framing of the policy topics at the time they appeared on the agenda.

The datasets were constructed by trained human coders applying consistent definitions of the content categories across the entire historical period and across all the data sources. Coding was done by at least two coders independently and checked for intercoder reliability (minimal 95 percent on main topic and 80 percent on subtopic level). After the training phase we had weekly or bi-weekly meetings to discuss the problematic cases. Consistency with other Agendas Projects was achieved through frequent contact with other project leaders (especially Denmark, France, and Spain).

Questions from parliament to a minister (vragenuurtje). Every Tuesday afternoon at 2 p.m., individual members of parliament invite ministers to the parliament to question them about a certain topic. Usually the MPs refer in their question to an issue that obtained media attention, which is a variable in the dataset. Please refer to Timmermans and Breeman (2010) when using this dataset.

Our data collection was in part funded by public organizations that requested that we should carry out commissioned research on attention patterns and their underlying mechanisms. In 2012, the Ministry of Internal Affairs asked us to look at how and why the attention to safety issues diversified in the past decades. As our graphs of two coalition agreements show, the variety of topics the governments linked to safety issues increased considerably between 1977 and 2010 (Figure 14.1).

An important finding in our study was that attention to safety-related events and incidents only leads to major policy plans if politicians and policymakers link these events and incidents to longer-term trends. In this way, we were able to distinguish two types of cascading effects: a short-term cascading effect over incidents and a long-term cascading effect resulting in long-term policy plans.

In the first type, cascading occurs when different actors such as journalists, politicians, and experts respond quickly to each other about an incident. Characteristic of this type is that incidents are considered as a “stand-alone” event with no link to other phenomena. When the incident is cleared, the media and all other actors lose interest and shift their attention to other matters. Thus in this process, attention is temporal.

The second type occurs when policymakers link different incidents with each other and consider these as part of a trend. A single incident does not usually lead to a major policy change. In our case studies of specific safety related issues we observed for instance how several youth crime incidents were linked to each other and were followed by a substantive policy program on immigrant integration. Once this policy program was put in place, incidents were more quickly connected to the observed trend and, as a self-reinforcing mechanism, these new incidents legitimized and strengthened the policy program. Thus, in this process, problem signals reinforced policy development.