We provide necessary and sufficient conditions for the identification (point-identification) of structural vector autoregressions (SVARs) with external instruments considering the case in which r instruments are used to identify g structural shocks of interest, r ≥ g ≥ 1. Novel frequentist estimation methods are discussed by considering both a partial shocks identification strategy, where only g structural shocks are of interest and are instrumented, and a full shocks identification strategy, where despite g structural shocks being instrumented, all n=g+(n-g) structural shocks of the system can be identified under certain conditions. The suggested approach is applied to investigate empirically whether financial and macroeconomic uncertainty can be approximated as exogenous drivers of US real economic activity, or rather as endogenous responses to first moment shocks, or both. We analyze whether the dynamic causal effects of nonuncertainty shocks on macroeconomic and financial uncertainty are significant in the period after the global financial crisis.
