A0543
Title: Averaging impulse responses
Authors: Christian Matthes - Indiana University (United States) [presenting]
Paul Ho - Federal Reserve Bank of Richmond (United States)
Thomas Lubik - Federal Reserve Bank of Richmond (United States)
Abstract: Impulse response analysis is a key tool for researchers to study the effects of shocks and policies on economic outcomes. When estimating impulse responses, economists have a wide range of options. For example, one can choose between local projections (LP) and vector autoregressions (VARs), Bayesian and frequentist methods, and different specifications. Each choice has its own drawbacks and benefits. It is well known that these choices can generate significantly different results. While there is a growing literature discussing conditions for which one approach might be preferred over another, in practical applications many of the conditions are likely to be difficult to verify. We propose the use of prediction pools to average impulse responses across different models.