Who cares about climate change? Analysing 20 years of plenary discussions with topic modelling

In the context of the past political, social, and most recent economic and geopolitical developments, energy policy has been the topic with the highest priority in the European Union (EU). Besides being one of the most critical political issues, energy policy is also inherently tied to climate change and sustainability. However, do EU Member States (MS) really prefer climate policies over other closely related yet often competing policy priorities of the Energy Union, the project launched in 2015? Thus, as energy and climate policy is perceived a highly sensitive area for many EU MS, following their different energy mixes and needs, my scholarly caution has been raised about the climate policy goals of the EU and their feasibility with other priorities, which aims at building a secure, competitive, and affordable energy market. In this paper, I investigated the policy preferences of the EU MS, performing a topic modelling analysis of more than twenty years of plenary debates in the European Parliament (EP). For this extensive data inquiry, I took advantage of an unprecedentedly large dataset of more than 34 500 energy-related speeches (consisting of more than 4 million words) addressed by Members of the EP (MEPs) during plenary debates. I intend to cast light on the underlying features of text reflecting two decades of energy policy debates in the EP and encode its structure into an estimation of categories (topics). Application of an unsupervised machine learning technique enabled me to identify the temporal changes in the organisation of the energy and climate policy debates and clearly showed evident regional cleavages to be of main "conflict" line structuring probability of addressing climate policy (or sustainability in general), i.e., offering to be of higher importance for EU MS from the area of Northwest Europe. In contrast, MEPs from Central and Eastern Europe and the Baltics are among the most frequent ones mentioning the security of supply issues. Against this outcome, this article also asks what lessons can be learnt from this for the EP's future role in adopting a more ambitious climate policy (as an effect of Russia's invasion of Ukraine) and its implementation. Besides, methodologically, this research also illustrates the potential and need for computational methods that enable the analysis of large datasets and complement conventional social science methods.