Generalized Covariance Estimator

We consider a class of semi-parametric dynamic models with strong white noise errors. This processes includes the standard Vector Autoregressive (VAR) model, nonfundamental structural VAR, mixed causal-noncausal models, as well nonlinear such (multivariate) ARCH-M model. For estimation in this class, we propose Generalized Covariance (GCov) estimator, which is obtained by minimizing residual-based multivariate portmanteau statistic an alternative to Method Moments. derive asymptotic properties GCov estimator and associated statistic. Moreover, show that estimators are semi-parametrically efficient statistics asymptotically chi-square distributed. The finite sample performance illustrated simulation study. also applied model cryptocurrency prices.