An empirical study on the influence of investor sentiment on the U.S. Stock Market: evidence based on minimum spanning trees and Multifractal Detrended Fluctuation Analysis

This thesis investigates the impact of investor sentiment on asset pricing dynamics in the U.S. stock market. Traditional asset pricing models have limitations in capturing investor sentiment. To address this issue, this research constructs two sentiment indices using error terms from the CAPM and FF5F models. The Minimum Spanning Tree Method is employed to investigate sentiment-based co-movements between stocks and Multifractals Analysis is used to study the inefficiency of the U.S. stock market. The empirical results reveal significant sentiment-driven co-movement patterns and multifractal characteristics that challenge the Efficient Market Hypothesis. Furthermore, the impact of the COVID-19 pandemic is examined, and it is found that co-movements increased during the pandemic, and the U.S. stock market was less efficient due to the pandemic. Overall, the findings underscore the importance of investor sentiment in understanding asset pricing dynamics and suggest the potential for enhanced asset pricing models incorporating sentiment-related information. This research may have significant implications for improved decision-making in investment, trading, and policymaking.