Statistics of Extreme Risks
When we model returns using a GARCH process with normally distributed innovations, we have already taken into account the second stylized fact (see Chapter 12). The distribution of the random returns automatically has a leptokurtosis and larger losses occurring more frequently than under the assumption that the returns are normally distributed. If one is interested in the 95%-VaR of liquid assets, this approach produces the most useful results. For the extreme risk quantiles such as the 99%-VaR and for riskier types of investments the risk is often underestimated when the innovations are assumed to be normally distributed, since a higher probability of particularly extreme losses than a GARCH process ε t with normally distributed Z t can produce.
KeywordsPareto Distribution Exceedance Probability Generalize Pareto Distribution Empirical Distribution Function Extremal Index
Unable to display preview. Download preview PDF.
- Artzner, P., Delbaen, F., Eber, F.-J. and Heath, D. (1997). Thinking coherently, Risk Magazine.Google Scholar
- Embrechts, P., Klüppelberg, C. and Mikosch, T. (1997). Modelling Extremal Events, Springer Verlag, Berlin.Google Scholar
- Franke, J., Härdle, W. and Kreiss, J. (2003). Nonparametric estimation in a stochastic volatility model, Recent advances and trends in Nonparametric Statistics.Google Scholar