Multi-step Forecasting with Large Vector Autoregressions
Essays in Honor of M. Hashem Pesaran: Prediction and Macro Modeling
ISBN: 978-1-80262-062-7, eISBN: 978-1-80262-061-0
Publication date: 18 January 2022
Abstract
This chapter investigates the performance of different dimension reduction approaches for large vector autoregressions in multi-step ahead forecasts. The authors consider factor augmented VAR models using principal components and partial least squares, random subset regression, random projection, random compression, and estimation via LASSO and Bayesian VAR. The authors compare the accuracy of iterated and direct multi-step point and density forecasts. The comparison is based on macroeconomic and financial variables from the FRED-MD data base. Our findings suggest that random subspace methods and LASSO estimation deliver the most precise forecasts.
Keywords
Acknowledgements
Acknowledgments
The authors would like to thank participants at the RMSE workshop and the ISCEF conference, in particular our discussant, Gilles Dufrénot, for valuable comments. Additionally, we thank Tom Boot, Agnieszka Markiewicz, Didier Nibbering, Richard Paap, and an anonymous referee. We thank SURFsara for access to the Lisa Computer Cluster. The opinions expressed are those of the authors and should not be attributed to DNB or Egeria.
Citation
Pick, A. and Carpay, M. (2022), "Multi-step Forecasting with Large Vector Autoregressions", Chudik, A., Hsiao, C. and Timmermann, A. (Ed.) Essays in Honor of M. Hashem Pesaran: Prediction and Macro Modeling (Advances in Econometrics, Vol. 43A), Emerald Publishing Limited, Leeds, pp. 73-98. https://doi.org/10.1108/S0731-90532021000043A005
Publisher
:Emerald Publishing Limited
Copyright © 2022 Andreas Pick and Matthijs Carpay