Title: A new approach to identify noise shocks
Authors: Luca Benati - University of Bern (Switzerland)
Joshua Chan - Australian National University (Australia)
Eric Eisenstat - The University of Queensland (Australia) [presenting]
Gary Koop - University of Strathclyde (United Kingdom)
Abstract: The idea that news about future productivity can play an important role in business cycle fluctuations infuses much current macroeconomic research. This paper addresses the question of whether the news is noisy, so that agents cannot quickly disentangle genuine news from a noisy signal. A logical implication of agents inability to distinguish news and noise shocks on impact is that the immediate response of the economy to the two shocks will be the same. We provide illustrations of this general property within several macroeconomic models. We then exploit this restriction, together with the fact that whereas news shocks portend future variation in the variable of interest, noise shocks do not, in order to identify news, noise and surprise TFP shocks within a structural VARMA framework. Our experiments show that sizeable systems of 8 to 15 variables are needed to identify these shocks and we develop Bayesian methods to estimate large, structural VARMAs. In an empirical application, evidence suggests that TFP noise shocks play a minor role in macroeconomic fluctuations, explaining negligible fractions of the forecast error variance of the main macroeconomic variables.