Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/300819 
Year of Publication: 
2024
Series/Report no.: 
DIW Discussion Papers No. 2095
Publisher: 
Deutsches Institut für Wirtschaftsforschung (DIW), Berlin
Abstract: 
The shocks in structural vector autoregressive (VAR) analysis are typically assumed to be instantaneously uncorrelated. This condition may easily be violated in proxy VAR models if more than one shock is identified by a proxy variable. Correlated shocks may be obtained even if the proxies are uncorrelated and satisfy the usual relevance and exogeneity conditions individually. Examples from the recent proxy VAR literature are presented. It is shown that assuming uncorrelated proxies that satisfy the usual relevance and exogeneity conditions individually actually over-identifies the shocks of interest and a Generalized Method of Moments (GMM) algorithm is proposed that ensures orthogonal shocks and provides efficient estimators of the structural parameters. It generalizes an earlier GMM proposal that works only if at least K - 1 shocks are identified by proxies in a VAR with K variables.
Subjects: 
Structural vector autoregression
proxy VAR
external instruments
correlated shocks
Generalized Method of Moments
JEL: 
C32
C36
E52
Document Type: 
Working Paper

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.