Proxy-identification of a structural MGARCH model for asset returns
Journal
Swiss Finance Institute Research Paper Series
Type
working paper
Date Issued
2024-10-17
Author(s)
Abstract
We extend the multivariate GARCH (MGARCH) specification for volatility modeling by developing a structural MGARCH model that targets the identification of shocks and volatility spillovers in a speculative return system. Similarly to the proxy-SVAR framework, we leverage auxiliary proxy variables to identify the underlying shock system. The estimation of structural parameters, including an orthogonal matrix, is achieved through techniques derived from Riemannian optimization. Our analysis of daily S&P 500 returns, 10-year Treasury yields, and the U.S. Dollar Index, employing news-driven instrument variables, identifies an equity and a bond market shock.
Funding(s)
Language
English
Keywords
identification
Riemannian optimization
structural MGARCH
structural modeling
variance decomposition
volatility spillovers
Pages
56
Division(s)