Ignace De Vos has published a new study in the Journal of Econometrics regarding the estimation of economic shocks.
In the article, "GLS estimation of local projections: Trading robustness for efficiency," Ignace De Vos and Gerdie Everaert examine how Generalized Least Squares (GLS) transformations affect the properties of local projection estimators of impulse responses. While local projections are valued for their robustness to model misspecification, they tend to exhibit relatively high variance. GLS transformations have been proposed to address this drawback. However, the authors show that such transformations either provide no efficiency gains or compromise the robustness to misspecification that makes local projections attractive. Their findings highlight that applying GLS transformations thus entails a trade-off between efficiency and robustness.