ESTIMATING VALUE AT RISK DURING THE STORM: A STUDY OF 25 EUROPEAN MARKETS
ESTIMATING VALUE AT RISK DURING THE STORM: A STUDY OF 25 EUROPEAN MARKETS

Summary/Abstract: This paper investigates which approach to estimating a common risk metric - the Value at Risk (VaR)– yields optimal results in times of significant market turbulence. To this end we leverage data on 25 European stock exchanges over a 15-year period ending in December 2020. Using data on the first 14 years, we estimate the non-parametric, the parametric Gaussian and the Cornish-Fisher versions of the VaR and compare those estimates to the actual realization in the last year of the period. A number of error metrics are consulted with both the mean absolute percentage error (MAPE) and the root mean squared error (RMSE) showing that a Gaussian parametric VaR yields the most accurate approximation to the actual value. Some implications of these results are outlined.

• Details
• Contents