The impact of economic sentiment on financial portfolios during the recent turmoil

This paper aims to assess the influence of economic sentiment on financial portfolios during the recent period of the Covid-19 pandemic and the Russia–Ukraine crisis. We follow a dictionary-based approach to compute sentiment indexes related to 13 US portfolios. Financial sentiment is extracted from a data set of 4 974 000 tweets covering the period from October 4, 2019 to March 23, 2022. For these 13 indexes, we compute their level and variation to capture the new information and we propose three empirical exercises. Using contemporaneous regressions, our results show that the impact is relatively pronounced for 11 out of the 13 portfolios. The unresponsive portfolios are those based on the utilities and consumer staples sectors. Transforming our indexes into first differences allows us to obtain predictive power. In this framework, the sentiment effect is strongest in the energy, industrial and consumer goods sectors. We conclude with a portfolio management analysis and suggest that incorporating our market sentiment indexes into portfolio construction results in an increase in the Sharpe ratio for the majority of the sectors.