Refining Set-Identification in VARs through Independence; Jonathan Wright (Johns Hopkins University)

Abstract

Identification in VARs has traditionally mainly relied on second moments. Some researchers have considered using higher moments as well, but there are concerns about whether that is robust. We propose refining existing identification schemes by intersecting the traditional confidence intervals with a region that rules out shocks whose higher moments significantly depart from independence. This approach does not assume that higher moments help with identification; it is robust to weak identification.  It works well in simulations, but yields only modest gains in empirical applications. These results vindicate a cautious approach to exploiting higher moments.