Working Paper: CEPR ID: DP9049
Authors: Michael R. Wickens
Abstract: We find that forecasts from DSGE models are not more accurate than either times series models or official forecasts, but neither are they any worse. We also find that all three types of forecast failed to predict the recession that started in 2007 and continued to forecast poorly even after the recession was known to have begun. We investigate why these results occur by examining the structure of the solution of DSGE models and compare this with pure time series models. We show that the main factor is the dynamic structure of DSGE models. Their backward-looking dynamics gives them a similar forecasting structure to time series models and their forward-looking dynamics, which consists of expected values of future exogenous variables, is difficult to forecast accurately. As a result we suggest that DSGE models should not be tested through their forecasting ability.
Keywords: DSGE models; forecasting; VAR models
JEL Codes: C5; E1
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| DSGE models' forecasting performance (E17) | time series models' forecasting performance (C53) |
| DSGE models' forecasting performance (E17) | official forecasts' forecasting performance (C53) |
| failure of forecasts to predict recession (F37) | structural dynamics of DSGE models (E13) |
| model structure of DSGE models (E13) | forecasting performance (C53) |
| accuracy of DSGE models' forecasts (E17) | forecasts of current and future exogenous variables (C53) |
| structural restrictions of DSGE models (E13) | forecasting performance (C53) |