|
Estimating Financial Risk under Time-Varying Extreme Return Behavior |
N. Wagner (2003): Operations Research Spectrum 25: 317-328
Potentially increasing volatility and downside risk is essential to
financial risk management which is concerned with the tails, or
particularly, the lower tail, of the distribution of speculative asset
returns. Applying extreme value theory, the present paper outlines a
simple model capturing time-varying tail behavior and studies
conditional daily return quantiles for the German DAX. Our results
indicate an overall
increased risk of large one-day holding period losses related to a
structural break given by the 1987 crash, systematic underestimation of
the magnitude of extreme quantiles as well as clustering in estimated
quantile exceedances that cannot be explained by the forecasting model.
|
