Working Paper: CEPR ID: DP8321
Authors: Sandra Eickmeier; Wolfgang Lemke; Massimiliano Marcellino
Abstract: We propose a classical approach to estimate factor-augmented vector autoregressive (FAVAR) models with time variation in the factor loadings, in the factor dynamics, and in the variance-covariance matrix of innovations. When the time-varying FAVAR is estimated using a large quarterly dataset of US variables from 1972 to 2007, the results indicate some changes in the factor dynamics, and more marked variation in the factors' shock volatility and their loading parameters. Forecasts from the time-varying FAVAR are more accurate than those from a constant parameter FAVAR for most variables and horizons when computed in-sample, and for some variables in pseudo real time, mostly financial and credit variables. Finally, we use the time-varying FAVAR to assess how monetary transmission to the economy has changed. We find substantial time variation in the volatility of monetary policy shocks, and we observe that the reaction of GDP, the GDP deflator, inflation expectations and long-term interest rates to a same-sized monetary policy shock has decreased since the early-1980s.
Keywords: FAVAR; Forecasting; Monetary Transmission; Time-Varying Parameters
JEL Codes: C3; C53; E52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Monetary policy shocks (E39) | Volatility of monetary policy shocks (E39) |
| Monetary policy shocks (E39) | GDP (E20) |
| Monetary policy shocks (E39) | GDP deflator (E31) |
| Monetary policy shocks (E39) | Inflation expectations (E31) |
| Monetary policy shocks (E39) | Long-term interest rates (E43) |
| Monetary policy shocks (E39) | Transmission mechanism (F42) |