We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Daily Volatility Forecasts: Reassessing the Performance of GARCH Models.
- Authors
McMillan, David G.; Speight, Alan E. H.
- Abstract
The article presents a forecast on the performance of GARCH model in Economics. The accurate estimation and forecasting of volatility in financial markets is an essential issue. Variations in market returns and other economy-wide risk factors are a main feature of asset and portfolio management and play a key role in derivatives pricing models. Moreover, price movements are linked to the arrival of news at an intra-day level. Recent empirical research has suggested that despite its success in modeling temporal dependence in conditional variance, the GARCH model is unable to provide accurate ex post volatility measures, such that simple statistical models, for example historical volatility and moving average models, provide superior forecast measures. This conclusion has been supported, to a greater or lesser extent, using data for several international stock markets and exchange rates. However, a new line of enquiry has suggested that the failure of GARCH models to produce accurate forecasts is not due to a failure of the model, but to the inappropriate use of ex-post squared returns of the same frequencies as the forecast evaluation measure of true volatility.
- Publication
Journal of Forecasting, 2004, Vol 23, Issue 6, p449
- ISSN
0277-6693
- Publication type
Academic Journal
- DOI
10.1002/for.926