Civil Society, Media, and Policy Outputs: Civil society events’ effects on environmental protection expenditure in Europe

Political scientist and sociologists alike have long considered movement politics to be an important determinant of policy changes as an alternative channel through which grievances are expressed. However, while this topic has been under scrutiny for almost four decades the progress made was mostly unsatisfactory (Amenta et al. 2010). On the one hand, the literature suffers from conceptualization and measurement problems regarding policy impact. On the other, there is also a need to extend the empirical scope of previous analyses by covering a larger sample of countries, more than a single causal factor, and considering the interactive and contingent effects of these factors. Recognizing these needs in the study of movements’ consequences, this paper analyzes the effect of civil society events on environmental expenditure in 15 European countries between 1996 and 2013 using GDELT event data and EUROSTAT expenditure data. The first part of the paper is dedicated to providing a thorough conceptualization of what policy impact means. The conceptualization furthermore involves demarcation from related terms such as success, outcomes, or outputs that are sometimes discussed under the umbrella term impact. The second part of the paper identifies several theoretical models put forward in the literature on movements’ impact and tests their related hypotheses using pooled time series analyses. Firstly, we test the information-and-resource hypothesis according to which civil society activities provide information necessary for reelection of politicians who, therefore, tend to listen to it (Soule and Olzak 2004, Gillion 2014). The yearly number of environmental civil society events, their conflictual nature, and their media coverage is used for this purpose. Secondly, the political opportunities hypothesis is included by using a measure of the openness of the electoral system (Kitschelt 1986, Kriesi 2004). Thirdly, political mediation models (Amenta et al. 1994, 1996) argue that successful mobilization requires mediation by supportive or sympathetic actors in political institutions and, thus, a measure of supportive parties in Parliament is tested. Finally, the paper also considers the public opinion hypothesis (Burstein and Linton 2002) and the framing hypothesis by looking at public opinion support and media tone of events. By integrating the different models of the impact that movements have on policy into a more comprehensive analysis and by recognizing the different channels though which citizens affect policy, this paper aims to advance and bridge the studies of movement politics and of political representation. While the findings of this paper are limited to one issue area due to the difficulty in finding matching participation and policy output data, they set the first step towards a cross-issue design that could also shed light on how these models might play out in different issue areas.