VaR and CVaR of Czech Financial Assets Returns Using GARCH Models with Heavy Tails Distributions

Value at Risk (VaR) and Conditional Value at Risk (CVaR) are popular measures used to estimate the risk exposure of investments of risky financial assets. Their accurate evaluation is important as they affect the further actions in risk management. As volatility of returns of financial assets exhibit heteroskedastic behavior, GARCH models are often used to capture this property. Further, it is also known that returns of financial assets are often distributed with heavy tails. So far this character is modeled with heavy-tailed distributions. We investigate the ability of the most often used heavy tailed distributions in GARCH and GJR-GARCH models to evaluate how well they can help to properly compute VaR and CVaR on several types of Czech financial time series as index PX, CEZ stock price and exchange rate EURCZK. In the conclusion we offer some inferences for practical implications for the use of heavy-tailed distributions in GARCH models for the stated purpose.