Abstract
We assess the role of monetary policy news shocks in the context of a medium scale DSGE model estimated on US data. We estimate several versions of the model and find decisive evidence in favour of the inclusion of monetary policy news shocks over a two-quarter horizon. According to our results, monetary policy news shocks account for a non-negligible fraction of the variance of real variables, especially at shorter forecast horizons. Further, we document that the importance of monetary policy news shocks goes beyond what was observed in recent years. The historical importance of monetary policy news shocks dates back to the 1999–2006 period when the official FOMC statements provided information about both the current policy setting and the expected future policy path. We also show that adding monetary policy news shocks to the model does not lead to identification problems.
Citation
@article{ISKREV2017108,
title = {Monetary policy shocks: We got news!},
journal = {Journal of Economic Dynamics and Control},
volume = {74},
pages = {108-128},
year = {2017},
issn = {0165-1889},
doi = {https://doi.org/10.1016/j.jedc.2016.10.010},
url = {https://www.sciencedirect.com/science/article/pii/S0165188916301749},
author = {Sandra Gomes and Nikolay Iskrev and Caterina Mendicino},
keywords = {DSGE models, Bayesian estimation, News shocks, Local identification, Business cycles, Forward guidance},
abstract = {We assess the role of monetary policy news shocks in the context of a medium scale DSGE model estimated on US data. We estimate several versions of the model and find decisive evidence in favour of the inclusion of monetary policy news shocks over a two-quarter horizon. According to our results, monetary policy news shocks account for a non-negligible fraction of the variance of real variables, especially at shorter forecast horizons. Further, we document that the importance of monetary policy news shocks goes beyond what was observed in recent years. The historical importance of monetary policy news shocks dates back to the 1999–2006 period when the official FOMC statements provided information about both the current policy setting and the expected future policy path. We also show that adding monetary policy news shocks to the model does not lead to identification problems.}
}