Working Paper: CEPR ID: DP14140
Authors: Adrian Pagan; Michael R. Wickens
Abstract: This paper discusses, and provides new evidence on, the view that "theory, while essential, should be regarded as a flexible framework rather than a straightjacket, because features that the theory abstracts from may be important in practice". It considers how best to assess the empirical performance of tightly specified models such as DSGE models, and loosely specified models such as SVARs. These issues are illustrated using various New Keynesian models. We conclude that the challenge is to incorporate flexibility into the theory in such a way as to be compatible with both the theory and the data.
Keywords: Model Testing; DSGE Models; SVARs
JEL Codes: C52; E3
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| theoretical frameworks (B54) | empirical performance of macroeconomic models (E17) |
| traditional DSGE models (E13) | fail to capture important dynamics in the data (C55) |
| traditional measures (like MDD) (C52) | inadequately capture model performance (C52) |
| assumption of uncorrelated shocks (C29) | affect validity of standard procedures (like variance decomposition) (C20) |
| strong priors (D80) | impose undue restrictions on estimates (C51) |