Extreme Value Theory (EVT) methods are used to investigate the asymptotic distribution of the lower tail for daily returns in the Athens Stock Exchange (ASE) over the period 1986 to 2001. Overall, the Generalised Logistic (GL) distribution is found to provide adequate descriptions of the stochastic behaviour of the ASE index extreme minima over the period studied. However, using moving windows techniques we show that the parameters of this distribution appear to vary with a tendency to become less fat tailed over time. This paper argues that market risk measurement models that are able to exploit this time varying behaviour could lead to more accurate risk estimates and therefore, have potentially important implications for risk assessment.
- Extreme value theory
- Probability weighted moments
- Anderson-Darling goodness of fit test
- Generalised extreme value distribution
- Generalised logistic distribution