Abstract

Samar Hossan El Din
ANALYTICAL STUDY OF NEWS IMPACT ON CONDITIONAL VOLATILITY OF EGX 30 (1998-2012) BEFORE and AFTER JANUARY 25TH REVOLUTION
Since the early 1980s, there has been a movement toward incorporating more behavioural science into finance in US and UK academic research. However, this area is in its embryonic stage in the Middle East. This is the reason behind the originality of this research. It is applied on the Egyptian stock market, and the sample examined consists of EGX 30 daily log-returns along fourteen years, where full sample is divided into three subsamples to be tested. The subsamples are chosen based on volatility levels and the breakpoint caused by January 25th revolution supported by Chow breakpoint test. Investors' behaviour is modelled based on Shiller-Sentana-Wadhwani theoretical model and TGARCH and EGARCH econometric models. It is found that news impact differs between high and low volatility samples. As volatility persistence is high, bad news has greater impact on it, while as volatility persistence is relatively low, bad news impact dies away by time. Concerning returns error autocorrelation, low volatility samples, returns error autocorrelation is positive but differs in its significance and implies presence of negative feedback trading effect, while results for high volatility sample are inconclusive.