Type
Text
Type
Dissertation
Advisor
Rachev, Svetlozar | Chen, Xinyun | Kim, Aaron | Xiao, Keli | Djuric, Petar.
Date
2015-12-01
Keywords
Applied mathematics | Back testing, Dynamic volatility process, Regime Switching model
Department
Department of Applied Mathematics and Statistics.
Language
en_US
Source
This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree.
Identifier
http://hdl.handle.net/11401/76549
Publisher
The Graduate School, Stony Brook University: Stony Brook, NY.
Format
application/pdf
Abstract
This thesis is to incorporate Markov Regime Switching model with Fractionally Integrated process in order to capture abrupt change and regime persistence simultaneously in long memory dynamic volatility process. We adapt truncated ARCH\(\infty\) to estimation scheme of our model. We carry out standard back testing procedure to validate Regime Switching FIGARCH VaR based forecasts, on S\&P 500 and SHSZ 300 data in 1 minute and 5 minute frequencies. In the Chapter 1, Regime Switching model , and parameter estimation steps based on truncated ARCH infinite and Hamilton filter will be given. Topics like stationary conditions of RS-FIGARCH and standard Normal Tempered Stable distributions as fat-tailed innovations of time series are also covered. In Chapter 2, Fractionally Integrated GARCH is reviewed, and incorporated with Regime switching model. Modified likelihood ratio based test proposed by \cite{Kasahara2013} is introduced as test against multiple regimes. Chapter 3 is to discuss VaR-based back testing procedure, using China's ShanghaiShenzhen 300 Index log return series and S\&P 500 log return series, both in 1 minute and 5 minute frequencies. Backtesing results are given, 99\%, 99.5\% and 99.9\% VaRs are compared with log returns illustratively and violations of Kupiec test at 0.01 and 0.05 significance levels are present as well. We claim that Regime switching FIGARCH not only can be used in risk management but has potential to be used in portfolio optimization. | 86 pages
Recommended Citation
Zhang, Xiao, "Regime switching and long memory in High Frequency financial data" (2015). Stony Brook Theses and Dissertations Collection, 2006-2020 (closed to submissions). 2449.
https://commons.library.stonybrook.edu/stony-brook-theses-and-dissertations-collection/2449