The paper makes a critical assessment of the Principal Components-GARCH (PCGARCH) model and argues why, when dealing with hundreds or thousands of variables, this model comes up as the most appropriate to be used. The suitability originates from the perspective of quality/cost ratio of volatility forecasts, allowing for a trade-off between quality and costs when computational efforts are significant. PCGARCH not only provides a method that allows for simpler volatility modeling, reducing significantly the computational time and getting rid of any problem that may arise from complex data manipulations, but also improves the modeling process quality by ensuring a stricter control of noise due to more stable correlation estimates.

Item Type:	Refereed Article
Keywords:	GARCH models, volatility forecasting, econometric models, evaluating forecasts, nonlinear time series
Research Division:	Commerce, Management, Tourism and Services
Research Group:	Banking, finance and investment
Research Field:	Financial econometrics
Objective Division:	Economic Framework
Objective Group:	Measurement standards and calibration services
Objective Field:	Measurement standards and calibration services not elsewhere classified
UTAS Author:	Matei, M (Dr Marius Matei)
ID Code:	99868
Year Published:	2012
Web of Science® Times Cited:	1
Deposited By:	Economics and Finance
Deposited On:	2015-04-14
Last Modified:	2015-05-11
Downloads:	0