Template-Type: ReDIF-Paper 1.0
Series: Tinbergen Institute Discussion Papers
Creation-Date: 2026-08-28
Number: 26-061/III
Author-Name: Justus Holman
Author-Workplace-Name: Vrije Universiteit Amsterdam 
Author-Name: Andre Lucas
Author-Workplace-Name: Vrije Universiteit Amsterdam 
Author-Name: Anne Opschoor
Author-Workplace-Name: Vrije Universiteit Amsterdam 
Title: Composite Univariate Modeling of Realized Covariance Matrix Dynamics and Volatility-at-Risk
Abstract: We propose a new model for realized covariance matrix dynamics using a composite of univariate time series models for its realized eigenvalues, linked by a copula function. The dynamics of each eigenvalue are based on a conditional F distribution, thus
allowing for fat-tailedness and outliers in the realized covariance matrices. Given its composition from univariate elements, the static parameters of the new model can be estimated efficiently by maximum likelihood. In an empirical application, we show that the new model outperforms relevant recent benchmarks in an extensive portfolio (Conditional) Volatility-at-Risk (VolaR) application.
Classification-JEL: C22, C32, C58
Keywords: 
File-URL: https://papers.tinbergen.nl/26061.pdf
File-Format: application/pdf
File-Size: 2.237.214 bytes
Handle: RePEc:tin:wpaper:20260061