Working Paper: NBER ID: w27771
Authors: Michele Lenza; Giorgio E. Primiceri
Abstract: This paper illustrates how to handle a sequence of extreme observations—such as those recorded during the COVID-19 pandemic—when estimating a Vector Autoregression, which is the most popular time-series model in macroeconomics. Our results show that the ad-hoc strategy of dropping these observations may be acceptable for the purpose of parameter estimation. However, disregarding these recent data is inappropriate for forecasting the future evolution of the economy, because it vastly underestimates uncertainty.
Keywords: VAR; COVID-19; macroeconomics; shock volatility
JEL Codes: C11; C32; E32; E37
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Extreme observations during the COVID-19 pandemic (C46) | Estimation of VAR parameters (C51) |
| Treating extreme observations as conventional data (C55) | Misleading results in VAR estimation (C32) |
| Excluding pandemic data entirely (C80) | Preferable to including it without adjustments (Y20) |
| Proposed method with scaling factor for volatility (C58) | Impulse responses similar to pre-pandemic data (E32) |
| Methodology capturing increased uncertainty (C59) | Critical for forecasting during pandemic (C53) |