Abstract
We analyze total connectivity of volatility in the presence of factors structure in volatilies. Total connectivity of volatility can be captured by a VAR(1) model of log volatility. We estimate VAR(1) model coefficient Γ using Factor Adjusted Regularization Method (FARM) and the residual covariance matrix with Principal Orthogonal Complement Estimator (POET). Theoretically, we prove that the proposed estimator can estimate total connectivity consistently under the high-dimensional setting. Empirically, our estimator reveals interesting changes over a set of systemic risk events.
Post a Comment