Dynamic Returns Linkages and Volatility Transmission Between South African and World Major Stock Markets

Abstract

This paper analyses returns and volatility linkages between the South African (SA) equity market and the world major equity markets using daily data for the period 1995-2007. Also analysed is the nature of volatility, the long term trend of volatility and the risk-premium hypothesis. The univariate GARCH and multivariate Vector Autoregressive models are used. Results show that both returns and volatility linkages exist between the SA and the major world stock markets, with Australia, China and the US showing most influence on SA returns and volatility. Volatility was found to be inherently asymmetric but reasonably stable over time in all the stock markets studied, and no significant evidence was found in support for the risk-premium hypothesis.

Notes

1 For a description of this methodology see Lee, 2001.

2 The stock market exuberance index is derived from the standard portfolio arbitrage relationship. For a comprehensive discussion, derivation and computation of stock market exuberance index see Cifarelli and Paladino (2005:416-417).

3 For a description and application of the VAR methodology, see among others, Sims, 1980; Brooks, 2003; Lamba and Otchere, 2002.

4 Brooks (2002:469) among others suggests that equity returns exhibit asymmetric responses of volatility to positive and negative shocks which are attributed to leverage effects. Leverage effects is a situation whereby a fall in the value of a firm’s stock causes the firm’s debt to equity ratio to rise, which leads ordinary shareholders to perceive their future cash flow stream as being relatively more risky.

5 The difference between γ > 0 and γ ≠ 0 is that in the former case the parameter γ only takes a positive value and such an instance would imply that there is evidence for both leverage and asymmetric effects. In the latter case γ can take both positive and negative values. Should it take a negative value, then only evidence of asymmetric effects and not leverage effects exist in the data (Eviews 5, 2004:597).

6 The term Granger causality is used to distinguish between statistical causality, which will be investigated here using the VAR and real causality.