Impulse Responses by Pseudo-Panel Local Projections: An R Package

Abstract

The estimation of impulse responses to shocks across multiple units (countries/regions/firms, etc.) is usually of interest in the international economics literature. One commonly faced challenge is that the limited time series dimension of the data undermines the precision of unit-specific impulse responses. Researchers often resort to panel regression estimation to exploit cross-sectional variation and shrink standard errors. However, the extensive homogeneity restrictions imposed by panel regressions can usually not be justified. This paper attempts to generalize, extend, and build an R package for a strategy described in Berg, Curtis, and Mark (2023) to shrink unit-specific impulse response standard errors in the context of local projection estimations.