Key points are not available for this paper at this time.
This paper develops a set of innovative text-based indices capturing oil market sentiment and oil price uncertainty. The textual analysis includes over 6 million news items spanning from January 1982 to June 2021. The evidence shows that sentiment indicators readily react to economic and geopolitical events that impact oil prices, thereby enabling these indicators to predict oil prices accurately. In contrast, uncertainty measures have inherent weaknesses, thus yielding unreliable oil price forecasts. This research yields a novel and robust text indicator that provides valuable insights for predicting the intricate dynamics of crude oil prices, particularly excelling in short-term forecasts and during periods of economic recession.
Luigi Gifuni (Fri,) studied this question.