The question if women represent women in parliament is central for the evaluation of representation and policy interventions. In parliaments, politicians shape how men and women are represented in decision-making and public discussion. Gender differences at almost all stages of the policy-making process have been investigated. Research at the aggregate level has often focused on the link between women’s descriptive and their substantive representation, i.e., the presence of women in parliament and policy output that is congruent with women’s interests or policy preferences. We ask how women represent women across political topics in parliamentary speech.

We use a semi-supervised topic model trained on coded UK election manifestos to categorize text and estimate its ‘female focus’. Based on the empirical analysis of 1,200,000 speeches in the British House of Commons from 1989-2019, we use word embeddings to identify terms associated with women in common contexts. We develop an empirical measure of femininity to score parliamentary speeches. Based on this measure, we identify “female issues” inductively.

We show that while issues like social welfare, families and health are “feminin”, many economic and foreign affairs issues are too. We find that even after we control for age, party and topic, women still communicate more focused on women. Men on the other hand talk much more about male aspects, even inside issues that are considered “female”. This supports the argument that in parliamentary speech, women are more likely to address women concerns, even inside traditionally female issues and underlines the importance of descriptive representation.