Decomposition of Conditional Variance Persistence through Frequency-Domain Markov Chains

Stefano Mainardi

Studies in Economics and Econometrics, 2005, vol. 29, issue 3, 111-135

Abstract: A Markov-switching (MS) approach is applied here to identify news with different intensity, such as weak but persistent shocks and apparently stronger shocks whose effects turn out to taper off and fade quicker. Hypotheses, indications and inference problems of time-domain MS models are reviewed. A three-step procedure for volatility persistence decomposition is formulated. Based on ad hoc MS specifications and related conditional probabilities, the aim is to examine time series properties of variables in levels, identify cut-off points in the spectrum of an efficient estimator of volatility, and construct time-varying weights for a component GARCH. The procedure is applied to 1989-2004 daily prices of two commodities. Phases of positive price performance tend to be steeper, deeper and less persistent than the opposite occurrences. The cocoa price undergoes relatively higher and more unstable volatility persistence, but this result is partly influenced by the modelling of a fourth regime in the coffee price series. Variance decomposition does not prove to strengthen the forecasting capacity of GARCH modelling.

Date: 2005
References: Add references at CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/10800379.2005.12106395 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:rseexx:v:29:y:2005:i:3:p:111-135

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/rsee20

DOI: 10.1080/10800379.2005.12106395

Access Statistics for this article

Studies in Economics and Econometrics is currently edited by Willem Bester

More articles in Studies in Economics and Econometrics from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().