Uncertainty in "Straight Line"…Why not? A New Structural VARs Estimation and Identification Procedure

In: https://morethesis.unimore.it/theses/available/etd-03202019-132346/

Abstract

We propose a new structural VAR method to estimate uncertainty and identify its macroeconomic effects formalizing the idea that, if uncertainty affects macroeconomic variables linearly, then a suitable linear combination of these variables should reveal uncertainty. We define uncertainty as the conditional volatility of the 1-step-ahead structural shock, following a Jurado et al. (2015) regarding the idea that uncertainty is a conditional volatility; however, being interested in structural VAR methods, our focus is on the structural shocks. Focusing on a specific structural long-run shock and the related uncertainty, we propose a simple and coherent two-step econometric procedure to estimate long-run uncertainty and its effects on macroeconomic variables. Our measure of uncertainty turns out to be highly correlated with respect to existing measures of uncertainty. Moreover, we see that the uncertainty peaks in quarters characterized by important recognizable economic, institutional and political events, such as the first oil shock, the monetary policy shocks of the Volcker era and the Lehman Brother bankruptcy. Thus emerges the ability of the uncertainty shock to lead the economic cycle, causing negative effects on economic activity that will be overcome only in the medium run.