Investor Behaviour: An Examination of Investor Sentiment and Cognitive Dissonance

This thesis seeks to examine the roles of investor sentiment and cognitive dissonance on investor behaviour. The objectives of this thesis are: first, to investigate the impact of the interaction of investor sentiment with culture on momentum and post-earnings-announcement-drift by way of cognitive dissonance in international markets; second, using investor sentiment and analyst recommendations to examine how cognitive dissonance affects institutional herding in the U.S. financial market.

The effect of investor sentiment, culture as well as cognitive dissonance is examined for the two anomalies, momentum and post-earnings-announcement-drift. The investigation is carried out both across a wide range of countries and in two distinct culture groups. We investigate these issues by building on a specific behavioural model and by bringing together arguments from psychology and the cross-culture literature in relation to investor sentiment, culture and the notion of cognitive dissonance. We propose that cognitive dissonance will be evident when private or public news contradicts investors’ sentiment. This will cause a slow diffusion of such news being incorporated into stock prices, resulting in return continuation and people in different cultures experiencing different degrees of cognitive dissonance and in different situations. The empirical findings suggest that cognitive dissonance is a key driver in explaining these two anomalies across countries and in the two distinct cultures.

The interaction of investor sentiment and analyst recommendations on institutional herding is investigated by using two commonly used herding measures in the micro-level in the U.S. It suggests that cognitive dissonance is an important driver for institutional herding by taking account of the interaction between the two factors. Cognitive dissonance will be evident when analyst recommendation revisions conflict with sentiment, causing institutions to herd differently in the current and subsequent periods. The two herding measures allow us to capture different aspects of herding in the two periods and to gain better insights into spurious and intentional herding.