Leveraging unstructured data sources in asset pricing

Résumé

This thesis analyzes the impact on various aspects of asset valuation of the recent technological advances and the vast amount of data available today. Specifically, Chapter 2 frames the current state of the discipline in the new context generated by the latest technological disruptions. Next, Chapter 3 conducts a systematic literature review to identify an unexplored gap at the intersection among sustainability, asset valuation, and modern techniques such as Natural Language Processing (NLP). Subsequently, chapter 4 delves empirically into this gap, producing sustainability metrics for US companies with NLP and demonstrating their convenience when included in the Fama-French asset valuation model. Finally, Chapter 5 contributes to the persistent debate on the financial analysts’ predictive capability by comparing oil price forecasts with futures contracts using unstructured data.