The term structure of investor sentiment and stock return

Articles in Press

Document Type : Research Paper

Authors

1 Department of Finance and Insurance, Faculty of Management and Accounting, Shahid Beheshti University, Tehran, Iran

2 statistics Department, faculty of mathematics, Shahid beheshti university, Iran

Abstract

Objective: How Investor Sentiments impacts excess returns has always been a topic of interest in behavioral finance. Recently more studies have emphasized on the effect of investor sentiment on short term decision making. These studies demonstrate that irrational factors have more significant effect in short term decision making. This raises the question of whether investor sentiment exerts different impact on stock excess returns due to different time term. In this paper, the different effects of investor sentiment on returns are investigated in three short term frequencies (daily weekly and monthly). Next the impact of investor sentiment on returns is compared.

Method: In this paper, the investor sentiment index is constructed by the two methods of Kalman filter, and principal component analysis. Next the effect of investor sentiments combined with market premium factor, size and value factors on excess return is empirically examined through panel data regression in daily, weekly and monthly frequencies. At last, the betta coefficients of investor sentiment in the three frequencies are compared with each other.

Results: The results show that the investor sentiment index constructed with both kalman filter and principal component analysis methods has a significant and positive effect on excess returns in all three frequencies. Also, the results demonstrate that the effect of investor sentiment on returns is a monotonous decreasing function of time term. Also, the results for both Kalman filter and principal component analysis are consistent through all three frequencies.

Keywords

Main Subjects

Journal of Asset Management and Financing

Articles in Press, Accepted Manuscript
Available Online from 13 October 2025

Files

Share

How to cite

Statistics