I examine representative claims made by European Commissioners between 1985 and 2024. Drawing on a corpus of 23,057 speeches by European Commissioners, I analyse who is represented and how the represented are constructed. Using the Representative Claims Analysis (RCA) framework, I identify statements in which Commissioners claim to speak for specific social groups, interests, or values. The identification process is semi-automated, combining manual coding with BERT-based language models to detect representative claims at the sentence level. To map the structure of representation, I employ a bipartite network analysis linking Commissioners to the social groups and topics they invoke. Building on this, I apply causal network analysis to estimate how the gender of representatives and those they represent shapes claim-making behaviour. I further simulate counterfactual patterns of degendered representation. This multi-method design enables me to examine how representative behaviour varies across portfolios, time, and gender. My findings advance understanding of how representation is constructed within the EU’s executive branch, contributing to debates on the EU’s democratic legitimacy, the genderedness of European institutions, and the social representativeness of European governance. Further, I am able to show the usefulness of causal network analysis and counterfactual representation networks to assess the influence of a representative’s gender on their representational performances.
