Identification of investment strategies and risk premium in Stock Exchanges.

Abstract

Efficient market hypothesis (hereafter, EMH) has been widely discussed for over thirty years or so however it failed to explain various economic events observed in world financial markets. Therefore, Behavioural Finance goes beyond the limitations of the forementioned theory and with the use mainly of cognitive psychology, attempts to explain heterogeneous investor reactions and their consequences in real equity markets. There can be no doubt that some investors try to unveil trends in past stock prices and base their portfolio decisions on the expectation that these trends will remain constant throughout time. These investors are often found in literature as ‘feedback traders’. Yet, they exhibit positive feedback trading behaviour when they ‘buy low and sell high’ creating a particular pattern in the market and driving prices away from fundamentals. Following that, they formulate price bubbles, market crashes and high return volatility of asset prices is being observed. The aim of this study is to examine the trading strategies that lie behind the interactions of agents in two major U.S. indices – the Dow Jones and S&P 500. The research is conducted using daily stock prices from 02/01/1957 to 22/12/2001. The relevant time series are all examined for unit root using the Augmented Dickey-Fuller (1979) test while heteroscedasticity is appropriately taken into account with the use of a (G)ARCH-M model. For the first time, we implement a newly developed model, called – GARCH-M-MG by providing two fundamental variables; the risk premium and feedback trading captured under the non-linear dynamics of the model. Since the studied period is too long, we chose to subdivide it into two parts, with the break-even point dated back to 10/03/2000 when the Dot-Com bubble burst. Our findings reveal the following: - negative sign of risk premium that is mainly attributed to various ‘anomalies’ detected in the 3-month treasury bill determination. Normally we should expect a positive sign (according to the theory). - Even if positive feedback is evident and statistically significant, it does not affect the behaviour of risk premium.