Using Unconventional Parametric VaR Models to Evaluate Extreme Risk

Kuzman, Boris and Radosavljević, Katica and Puškarić, Anton (2026) Using Unconventional Parametric VaR Models to Evaluate Extreme Risk. International Review, 15 (1-2). pp. 95-108. ISSN 2217-9739

Text Kuzman, Radosavljević, Puškarić - International Review, 15(1-2) 2026.pdf - Published Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (551kB)

Abstract

This paper aims to mitigate the extreme risk of the German DAX index by constructing parametric multivariate VaR portfolios that combine the DAX with emerging market indices from East Asia, the MENA region, and Central and Eastern Europe. Extreme risk is evaluated using several parametric Value-at-Risk models, including the normal, logistic, hyperbolic secant, and Laplace distributions. Results show substantial variation across models: the normal VaR produces the lowest risk estimates, while the Laplace VaR yields the highest. Kupiec test results indicate poor performance of the normal VaR, whereas heavier-tailed distributions perform better, particularly in the MENA portfolio.

Actions (login required)

View Item