v9i4 vonBodman

This study explores the impact of IPO prospectus language on the prevalent underpricing in European IPOs, using natural language processing techniques. Specifically, it investigates whether a relationship exists between litigious, negative, positive, and uncertain language, as well as the degree of document similarity and IPO underpricing. For this purpose, qualitative text data is converted into quantifiable metrics using modern analysis techniques. The study presents new methodological approaches to textual analysis. The results establish a clear relationship between underpricing and multiple dimensions of prospectus language and highlights unique features of European markets. These include specific disclosure obligations of various market segments and the different listing types available to issuing firms. The results of the variables related to sentiment analysis all reveal significant relationships. However, no robust evidence emerges for variables related to document similarity. Overall, the introduced methodological approaches offer enhanced explanatory power over traditional methods, effectively contributing to the explanation of the underpricing phenomenon in European markets.