Applying Quantitative Text Analysis on the Study of EU Environmental Coalitions

Environmental regulation is a core responsibility of EU countries and the onus for the mitigation of and adaptation to climate change has emerged as a central task of the EU. The decision-making process on environmental issues, however, is strongly influenced by interest groups. The public perception of this state-to-interest-group-relationship often corresponds to a simple dichotomy: non-profit organisations versus business groups. From this perspective, environmental pressure groups advocate for stronger regulation of harmful practices and highlight the external effects of environmental degradation across issues like air pollution, climate change or biodiversity. Business groups are often associated with the opposing side, lobbying for less regulation and intervention into market transactions. Environmental issue coalitions are perceived to be monolithic and stable across time and issues. We want to know if that is really the case? This paper analyses the formation and stability of interest coalitions within EU environmental policy-making. The paper is inductive and explorative in nature. We employ innovative approaches of quantitative text analysis on large datasets in order to show how ideational coalitions evolve among interest and business groups. The analysis is based on EU consultations on environmental topics including renewable energy, climate change and biodiversity. Although EU public consultations represent a central engagement mechanism within EU governance, this aspect has not received much attention from the academic community. By focusing on stability and change over time and issue, we can track the spatio-temporal dynamics of ideational environmental policy coalitions in the EU consultation process. The main contribution of the paper is the integration of innovations in machine learning for text analysis with the study of network governance of EU environmental policy-making.