This study aims at comparing the variance structure of high (daily) and low (weekly, monthly) frequencies of data. By employing ARCH (1) and GARCH (1, 1) models, the study finds evidence that the intensity of the shocks are not equal for all the series. The study first finds that statistical properties of the three data series of returns are substantially different from one another and the persistence of conditional volatility is also different for the three series. The presence of persistency are more in the daily stock returns as compared to other data sets, which shows that the volatility models are sensitive to the frequencies of data series. In simple the results reveal that the variance structure of high frequency data is dissimilar from the low frequencies of data and variance structure in the daily data is much linked with the stylized facts associated with stock returns volatility.
Keywords: ARCH, GARCH models, KSE 100-index, persistence.
The most significant topic of research in the financial markets for the last three decades is the stock returns volatility. After the publication of Engle (1982) paper, on ARCH, the volatility has received considerable attention from researchers, practitioners and policy makers. This interest is due to the reason that the volatility is a risk measure and different participants use this for different reasons. The volatility is high for the developing and developed countries in recent years but is much higher for the developing countries. So volatility study is more important in the developing countries. After the crash of 1987, the need for volatility measurement is the focus of attention by the practitioners, regulatory concerns and empirical researchers.
Persistency in volatility clustering is normally due to the inefficiency in the market. Rizwan and Khan (2007) studied the volatility of the Pakistani stock market and found volatility clustering which signifies inefficiency in the stock market. They found that lagged returns are significant in explaining current returns. The volatility persistence measures the time period for which any shock has significant impact on variance.
Volatility has different phenomenon when it is measured on short, medium and long term basis. So the different frequencies must be examined to see the short, medium and long term affects of the volatility. Dawood (2007) investigated volatility in the Karachi stock exchange and found that in 1990’s the market has become more volatile both on short term (daily) and medium term (monthly) basis. He found that stock market reacts too actively to economic shocks but these reaction take place on daily basis and die away within a month.
High frequency data series is considered to be the most volatile series than the low frequencies of data. As Chang (2006) investigated the mean reversion behavior of different series...