OVERCONFIDENCE BIAS OF INVESTORS’ INVESTMENT DECISIONS ON GHANA STOCK EXCHANGE

ABSTRACT

The Efficient Market Hypothesis (EMH) postulates that all publicly available information on any firm’s stock future performance is incorporated in the firm’s stock price such that the actual price of the stock is equal to its intrinsic value. The EMH is argued to lead to “correct prices” which creates the necessary environment for efficient allocation of economic productive scarce resources. The documentation of empirical evidence of weak-form inefficiency on most stock markets especially Africa, has received much attention by regulators, researchers, investors and many other players in the financial industry. The Ghana stock market, for instance, is documented in the literature to be weak-form inefficient. In this study, the researcher turns to behavioural finance to test the rationality of the investors on Ghana stock market since the rationality assumption is one of the pillars upon which market efficiency is built. To test the rationality of the market participants, overconfidence behavioural bias is used as a conditional rationality test proxy. The concentration on overconfidence bias does not suggest that overconfidence is the only bias worth considering but it is due to the fact that it is reported in the literature to be one of the robust psychological behavioural biases. Overconfidence is the act of having a mistaken assessment and believing in these assessments too strongly. To test the overconfidence bias, a market-wide Vector Autoregressive (VAR) model and its associated impulse response function is used to investigate the lead-lag relationship between market returns and market trading volume. Also, granger-causality test was performed to test the causality between returns and trading volumes. This study uses the Ghana Stock Exchange All-Share index (GSE-ASI) and the Ghana Stock Exchange Composite Index (GSE-CI) monthly returns data. In addition, the monthly trading volume, monthly returns volatility, interest rate and inflation data from 2000 to 2015 are used for the study. The findings reveal a significant impact of past market returns on current trading volumes. The impulse response function analysis also showed that the response of trading volume to a shock of market return residual remains highly positive and significant at lag one and two. The granger causality reveals unidirectional causality running from returns to trading volumes. The empirical findings finally demonstrate that market participants on GSE exhibit conditional irrationality (overconfidence bias as proxy) in their investment decisions.