most beautiful quant
Posted: June 17th, 2016, 5:12 pm
I read a master's thesis Value at Risk: GARCH vs. Stochastic Volatility Models: Empirical Study AbstractThe thesis compares GARCH volatility models and Stochastic Volatility (SV)models with Student's t distributed errors and its empirical forecasting performanceof Value at Risk on ve stock price indices: S&P, NASDAQ Composite,CAC, DAX and FTSE. It introduces in details the problem of SV models MaximumLikelihood examinations and suggests the newly developed approachof Ecient Importance Sampling (EIS). EIS is a procedure that provides anaccurate Monte Carlo evaluation of likelihood function which depends uponhigh-dimensional numerical integrals.Comparison analysis is divided into in-sample and out-of-sample forecastingperformance and evaluated using standard statistical probability backtestigmethods as conditional and unconditional coverage.Based on empirical analysis thesis shows that SV models can perform atleast as good as GARCH models if not superior in forecasting volatility andparametric VaR.JEL Classication F12, C22, C52, C53, G15Keywords VaR, GARCH, Stochastic Volatility, backtesting,conditional coverage, unconditional coverageand then googled the author. Her picture isat http://www.icpcp.eu/team/team-members/viktoria-tesarova . My goodness!