Use this URL to cite or link to this record in EThOS:
Title: Modelling intraday stock price dynamics on the Malaysian stock exchange
Author: Haniff, Mohd Nizal
ISNI:       0000 0004 2748 7111
Awarding Body: Cardiff University
Current Institution: Cardiff University
Date of Award: 2006
Availability of Full Text:
Access from EThOS:
Access from Institution:
The introduction of the Autoregressive Conditional Heteroskedasticity (ARCH) model in 1982 by Engle revolutionized the econometric treatment of volatility. The Generalized ARCH (GARCH) model and its variants have proved to be useful in capturing stylised facts about financial markets, which include volatility clustering, leptokurtosis in the distribution of returns, mean reversion tendencies and leverage effects. The Periodic GARCH (PGARCH) variants proposed by Bollerslev and Ghysels (1996), in particular, made it possible to explicitly incorporate the effects of periodicity in financial time series into the parameters of the volatility models. An investigation of return volatility using high frequency Kuala Lumpur Composite Index (KLCI) returns data shows that the intraday volatility pattern follows the double U-shaped pattern, which is consistent with the findings of other studies on markets that are closed during the lunch hour. The study also investigates the best technique for modelling and forecasting the intraday periodicity on the Kuala Lumpur Stock Exchange (KLSE), using both the jointly estimated and the two-step filtration approaches with different PGARCH structures. The results indicate that the PGARCH models produce superior model fit, better forecasting performances and superior forecast quality than the standard GARCH equivalents. However, the results suggest that Value-at-Risk (VaR) models, constructed from the PGARCH forecasts, produce poor results. This study also investigates the integrated realized volatility measure introduced by Andersen and Bollerslev (1998a), which can be constructed by summing up intraday squared returns. The results suggest that the daily integrated realized volatilities constructed using different intraday return sampling frequencies, produce superior forecasting performances for the GARCH models when compared with the results of the same models using the daily squared returns. The VaR models constructed from the GARCH forecasts and the Autoregressive and Moving Average (ARMA) forecasts appear to satisfy the requirements of the framework for interval forecast evaluation.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available