Armand, AlexJimenez, Bruno Aragon2023-04-212023-04-212022-01-132021-12http://hdl.handle.net/10362/151975This paper studies how politicians in the Spanish Congress take advantage of their visibility to improve the results of their party in local elections. It is a novel intersection of natural language processing and political economy. Using transcripts of congressional debates, I analyze how lexicon changes near to a local election date. There are significant results indicating a 20-30 percentage points increase in subjective discourse as well as 1-1.9 percentage points higher participation in debates before the election date. This behaviour is especially salient to politicians associated with a right ideology.engMachine learningRegression discontinuityNlpPolitical economyElectionsLobbyingSpainmaster thesis203221915