Comparison of Value-at-Risk using various empirical methods for the portfolios of BRICT and G-7 countries in the long run
diplomová práce (OBHÁJENO)
Zobrazit/ otevřít
Trvalý odkaz
http://hdl.handle.net/20.500.11956/34246Identifikátory
SIS: 92220
Kolekce
- Kvalifikační práce [18159]
Autor
Vedoucí práce
Oponent práce
Gapko, Petr
Fakulta / součást
Fakulta sociálních věd
Obor
Ekonomie a finance
Katedra / ústav / klinika
Institut ekonomických studií
Datum obhajoby
8. 9. 2010
Nakladatel
Univerzita Karlova, Fakulta sociálních vědJazyk
Angličtina
Známka
Velmi dobře
This master's thesis deals with Value-at-Risk for equity portfolios. The distribution of daily returns of equity returns is not perfectly normal. Therefore, the use of the Delta- Normal Value-at-Risk (VaR) method is misleading. Accuracy of estimation may turn out to be failure for portfolios to measure VaR time to time. Therefore, two further methods, Modified VaR and Filtered Historical Simulation, are used for VaR estimation. The former estimates using Cornish-Fisher (1937) expansion and then the latter estimates using autoregressive model for mean equation, EGARCH for volatility and Filtered Historical Simulation (FHS) for VaR estimation i.e. AR (1) - EGARCH (1,1) - FHS methods; and also the performance of both the VaR estimates with Delta- Normal VaR estimate are compared. Last but not the least the implementation of various methods are discussed and analyzed on the two passive historical index portfolios, which represent some of the most attractive financial markets in the world economy.
This master's thesis deals with Value-at-Risk for equity portfolios. The distribution of daily returns of equity returns is not perfectly normal. Therefore, the use of the Delta- Normal Value-at-Risk (VaR) method is misleading. Accuracy of estimation may turn out to be failure for portfolios to measure VaR time to time. Therefore, two further methods, Modified VaR and Filtered Historical Simulation, are used for VaR estimation. The former estimates using Cornish-Fisher (1937) expansion and then the latter estimates using autoregressive model for mean equation, EGARCH for volatility and Filtered Historical Simulation (FHS) for VaR estimation i.e. AR (1) - EGARCH (1,1) - FHS methods; and also the performance of both the VaR estimates with Delta- Normal VaR estimate are compared. Last but not the least the implementation of various methods are discussed and analyzed on the two passive historical index portfolios, which represent some of the most attractive financial markets in the world economy.