Investor sentiment and the cross section of stock returns: a natural language processing approach
Type
case review (law)
Date Issued
2023-12-17
Author(s)
Abstract
We investigate how investor sentiment a ects the cross-section of stock returns using a data-driven Natural Language Processing (NLP) methodology. Daily sentiment is derived from various text data sources, including newspaper headlines, tweets from Stocktwits, and earnings call transcripts. We apply a state-of-the-art NLP model for sentiment classi cation, with labels generated based on one-day ahead stock returns. The model's output can be
interpreted as a one-day ahead return forecast, which we utilize for conducting portfolio sorts. The contribution of our study is twofold: rst, we directly derive sentiment from text data, eliminating the reliance on proxies; second, our labels are generated through a data-driven process rather than human annotation.
interpreted as a one-day ahead return forecast, which we utilize for conducting portfolio sorts. The contribution of our study is twofold: rst, we directly derive sentiment from text data, eliminating the reliance on proxies; second, our labels are generated through a data-driven process rather than human annotation.
Language
English
HSG Classification
contribution to scientific community
Event Title
CFE 2023
Event Location
Berlin, Germany
Event Date
16.-18. December 2023
Subject(s)
Contact Email Address
jule.schuettler@unisg.ch