Estimation pratique de modèles de volatilité stochastique à haute dimension avec des applications à l’incertitude macroéconomique au Québec et au Canada

Stochastic covariances are critical for macroeconomic and financial modelling, particularly in capturing uncertainty and dynamic interdependencies. This study introduces the Dynamic Factor Augmented VAR with Higher-order Multivariate Stochastic Volatility (DFAVAR-HMSV) framework, along with a computationally efficient estimation methodology. The proposed model captures complex dynamic interdependencies, leverage effects, and higher-order persistence in volatility structures. Applying this framework to construct uncertainty indices for Canada and Québec, the study provides critical insights into regional and national macroeconomic dynamics.

[ - ]
[ + ]