State-Dependent Dynamics of Overconfidence in Frontier Equity Markets: A Transfer Entropy Approach from Bangladesh

The study investigates overconfidence bias in the Bangladesh equity market through the relationship between the market returns, and the trading volume in a nonlinear, information-theoretic model. Building upon the traditional literature on returns and volume, the study differentiates between the total market returns and unexpected market returns, the latter being the unexpected information shocks represented under the Market Index Model. Transfer Entropy with bootstrap inference is used to determine directional and asymmetric causality across various market states, including bullish, bearish, crisis, extended crisis, and COVID-19. The findings indicate that the total market returns give weak and inconsistent evidence of overconfidence, which is bi-directional but limited information flow. Conversely, unexpected market returns have a statistically significant directional effect on trading volume, which represents strong evidence of overconfidence. The results also reveal that overconfidence is conditional as it is stronger in normal and bullish market contexts, and weaker during times of crisis. Asymmetric analysis reveals that the overreaction of investors is more pronounced when the market trends are negative, implying that unexpected losses stimulate an amplified trading effect due to the feeling of mispricing and recovery hopes. The results have significant implications on market efficiency, investor behavior and regulatory policies to improve market stability and facilitate informed financial decision-making.