Abstract
This paper studies the existence and timing of bubbles in South Africa’s stock market. An empirical model of bubble formation is tested against three competing models of asset price returns that rule out the existence of bubbles. The model controls for nonlinearities inherent in asset price returns by allowing for the existence of multiple regimes. The bubble model fits the data better than the competing models and suggests that the formation and existence of periodically collapsing bubbles are a reality.
1. Introduction
Research on stock market bubbles will continue to have relevance as long as market risk and policy concerns about the formation of asset prices and arbitrage opportunities exist. By now it is generally agreed that stock market shocks can have adverse consequences – and that the global repercussions are significant. Systemic stock market price shocks imply that actual returns to investment and financial stability are all interwoven in a complex financial system – present in both stock price commonalities in an existing market (Greenwood, 2005) and stock markets across countries (Engle et al. 1990). If bubbles exist and are in the process of growing, they are not only of interest to the investor and policy maker about their existence, but also the implications that bubbles would have ifthey were to burst. The study of bubbles is even more complex since it is not even generally agreed upon that bubbles even exist. In fact if economic agents are completely rational and forward looking the underlying price of the bubble is the fundamental price – i.e. bubbles cannot exist. The complication becomes more nuanced once one distinguishes between different kinds of bubbles such as credit fuelled bubbles or bubbles formed due to irrational behavior. There are many competing theories on the formation and existence of bubbles that are complicated – this means that the empirical tools to test and identify these theories are just as complicated.
5. Conclusion
Some evidence points to the formation and existence of speculative bubbles in the South African stock exchange market as given by the Johannesburg All Share Index. Periods of seemingly excessive returns, relative to returns tied to fundamentals, and significant slowdowns can be related to the formation of bubbles. Three stylized or alternative models of stock price returns are tested against a model of periodically collapsing bubbles. The bubble model used in this paper outperforms other stylized models of stock price determination—suggesting that excessive returns, relative to returns that are linked to fundamentals, arise due to either leveraged positions or irrational investor behavior. While the model helps to identify bubbles in probabilistic terms it does not indicate what behavior, credit fuelled or herding behavior, has led to its formation. Thus,the current paper only helps policy makers and investors identify bubbles, but refrain from prescribing how to manage, if at all, the formation of bubbles in South Africa. A further study might provide more insights into this matter.
