Modeling Sentiment, Temporal Volatility and Excess Returns: Empirical Evidence from Segmented Stock Market

Abstract

This study examines the ability of investor sentiment to predict conditional volatility and excess returns at both aggregate market and industry level in Pakistani stock market. Following the top-down-approach, a broad band investor sentiment index for Pakistan has been developed to empirically test this issue. A significantly positive contemporaneous as well as negatively lagged effect of investor sentiment is found on excess returns at aggregate market and industry level. It has also been confirmed that bullish (bearish) sentiment increases (decreases) volatility which in-turn affect the mean variance relationship. However, the commonality of the effect of investor sentiment via conditional volatility has not been uniform across industries.

Keywords: Sentiment; volatility; emerging stock market; principal component

JEL Classification: A12, G14