CMStatistics 2021: Start Registration
View Submission - CFE
A1361
Title: The predictive content of the textual political polarity index: The case of Italian GDP Authors:  Alessandra Amendola - University of Salerno (Italy)
Walter Distaso - Imperial College London (United Kingdom)
Alessandro Grimaldi - University of Salerno (Italy) [presenting]
Abstract: A data-driven approach is proposed to derive a Textual Political Polarity Index (\textit{TPPI}) based on the analysis of the entire collection of the verbatim reports of the Italian ``\textit{Senate of the Republic}''. The procedure allows us to build a set of polarity indices reflecting the impact of political debate - as well as agreement/disagreement within parties' groups - on a specific economic variable over time. In order to assess such an impact, we perform predictive regressions on a chosen macroeconomic variable - namely, the yearly Italian GDP growth rate. Results point to a nontrivial predictive power of the proposed polarity indices, which (importantly) do not rely on a subjective choice of an affective lexicon.