Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/147117 
Year of Publication: 
2015
Citation: 
[Journal:] e-Finanse: Financial Internet Quarterly [ISSN:] 1734-039X [Volume:] 11 [Issue:] 1 [Publisher:] University of Information Technology and Management [Place:] Rzeszów [Year:] 2015 [Pages:] 32-43
Publisher: 
University of Information Technology and Management, Rzeszów
Abstract: 
We investigate several promising algorithms, proposed in literature, devised to detect sudden changes (structural breaks) in the volatility of financial time series. Comparative study of three techniques: ICSS, NPCPM and Cheng's algorithm is carried out via numerical simulation in the case of simulated T-GARCH models and two real series, namely German and US stock indices. Simulations show that the NPCPM algorithm is superior to ICSS because is not over-sensitive either to heavy tails of market returns or to their serial dependence. Some signals generated by ICSS are falsely classified as structural breaks in volatility, while Cheng's technique works well only when a single break occurs.
Subjects: 
volatility
structural breaks
financial time series
logarithmic returns
Threshold-GARCH model
JEL: 
C19
C22
C58
Persistent Identifier of the first edition: 
Document Type: 
Article

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.