Title: Investor sentiment metrics and stock market returns: a study of the causality relationship using VAR models

Authors: Dorsaf Ben Aissia; Nizar Neffati

Addresses: ISCAE, University of Manouba, Tunisia ' Higher Institute of Computer Science and Management, University of Kairouan, Tunisia

Abstract: This paper examines the causality relationship between investor sentiment metrics and stock market returns. It considers survey, market, and composite sentiment indexes. It also introduces a dummy variable detecting the effect of economic crisis and decomposes sentiment into rational and irrational components. It uses VAR models and Granger tests, estimates Impulse Reaction Functions (IRFs) of the non-expected movement in investor sentiment, and proposes a forecast error variance decomposition (FEVD) approach to emphasise the importance of these movements on variables of the VAR models. Based on US data (S&P 500, Dow Jones, and NASDAQ indexes) from July 1965 to December 2021, we find a negative and significant relationship between investor sentiment and stock returns. This relationship is primarily explained by the irrational component of sentiment. In addition, we find a bi-directional Granger causality between stock returns and investor sentiment. Still, the IRFs and the FEVD study confirm the superiority of the survey indexes over the market indexes.

Keywords: investor sentiment metrics; stock market returns; causality relationship; VAR models; Granger tests; impulse response function; forecast error variance decomposition; FEVD.

DOI: 10.1504/AJFA.2023.134710

American Journal of Finance and Accounting, 2023 Vol.7 No.2, pp.90 - 124

Received: 28 Jul 2022
Accepted: 25 Jun 2023
Published online: 07 Nov 2023 *