Working Paper: CEPR ID: DP7423
Authors: Karel Mertens; Morten O. Ravn
Abstract: Empirical estimates of the impact of government spending shocks disagree on central issues such as the size of output multipliers and the responses of consumption and the real wage. One explanation for the disagreement is that fiscal shocks are often anticipated. Due to misspecification of the information set, anticipation effects may invalidate SVAR estimates of impulse responses. We use economic theory to derive a fiscal SVAR estimator that is applicable when fiscal shocks are anticipated. We study its properties and apply it to US data. We fail to find evidence that anticipation effects overturn the existing findings from the fiscal SVAR literature.
Keywords: anticipation effects; fiscal policy; structural vector autoregressions
JEL Codes: C32; E20; E32; E62
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| anticipation effects (D84) | interpretation of impulse responses (C22) |
| unanticipated permanent increase in government spending (E62) | output (C67) |
| unanticipated permanent increase in government spending (E62) | consumption (E21) |
| anticipated increase in government spending (H59) | private consumption (D19) |