Lino Wehrheim: Insights into a DH/EH Project...

Insights into a DH/EH Project: More than a Feeling – Building a Newspaper-Based Sentiment Index for the Berlin Stock Exchange (1872–1930)

Economic expectations are not only influenced by ‘hard’ information such as future cash flows but also by ‘soft’ parameters such as moods, emotions, and feelings. In this project, following John Maynard Keynes, we take up the notion of ‘animal spirits’ by looking at an institution that is particularly notorious for its ‘irrational exuberances’ (R. Shiller): the stock market. More precisely, we investigate the role of sentiment for the Berlin stock exchange between 1872 and 1930, back then the leading German stock market. We use different computational sentiment analysis techniques (dictionaries, machine learning) to extract sentiment from a daily column published in the Berliner Börsen-Zeitung, which was the most important financial newspaper at the time. We propose a daily sentiment index which we assume to reflect market participants’ general emotional state. In combination with content analysis methods, such as topic modelling, and stock market data we investigate the relationship between market development, market sentiment, and media narratives at the Berlin stock exchange to address the following questions: How did historical experiences, such as wars and political or economic shocks, change sentiment and how did sentiment influence financial markets? Has this influence changed over time? Which role did sentiment play in the development of financial and economic crises, and which were the narratives that accompanied high stock market fluctuations?

Dr. Lino Wehrheim is a postdoctoral researcher at the History Department at Regensburg University (Germany), working on economic and media history with a focus on computational text analysis.