We propose a nonrecursive identification scheme for uncertainty shocks that exploits breaks in the volatility of macroeconomic variables and is novel in the literature on uncertainty. This approach allows us to simultaneously address two major questions in the empirical literature: Is uncertainty a cause or effect of decline in economic activity? Does the relationship between uncertainty and economic activity change across macroeconomic regimes? Results based on a small-scale vector autoregression with US monthly data suggest that (i) uncertainty is an exogenous source of decline of economic activity, and (ii) the effects of uncertainty shocks amplify in periods of economic and financial turmoil.
