Working Paper: NBER ID: w30971
Authors: James Cloyne; Oscar Jord; Alan M. Taylor
Abstract: An impulse response is the dynamic average effect of an intervention across horizons. We use the well-known Kitagawa-Blinder-Oaxaca decomposition to explore a response’s heterogeneity over time and over states of the economy. This can be implemented with a simple extension to the usual local projection specification that nevertheless keeps the model linear in parameters. Using our new decomposition-based approach, we show how to unpack heterogeneity in the fiscal multiplier, an object that at any point in time may depend on a number of potentially correlated factors, including existing economic conditions and the monetary response. In our application, the fiscal multiplier varies considerably with monetary policy: it can be as small as zero, or as large as 2, depending on the degree of monetary offset.
Keywords: Impulse Response; Fiscal Multiplier; Monetary Policy; State Dependence; Local Projections
JEL Codes: C54; C99; E32; E62; H20; H5
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Fiscal interventions (H39) | Economic outcomes (F69) |
| Policy interventions (D78) | Other variables impact outcomes (C39) |
| Improper identification of state variables (C32) | Biases in composition effect (D91) |
| Fiscal easing (E62) | Monetary tightening (E52) |
| Monetary policy environment (E52) | Variation in fiscal multipliers (E62) |