CONDITIONAL HETEROSCEDASTICITY IN TIME SERIES OF STOCK RETURNS: A REVISIT

Sinan Yildirim
Texas Wesleyan University
ABSTRACT
This study performs the tests reported in Akgiray (1989) by using more recent data set
over a longer time period in order to shed light on stock market volatility. The findings reveal that
time series of daily stock returns are significantly dependent. ARCH and GARCH models simulate
market volatility. Thus, time series behavior of stock market volatility can be modeled by
conditional heteroscedastic processes. When compared statistically, GARCH forecasts are better.