Working Paper: CEPR ID: DP6758
Authors: Kerstin Bernoth; Andrew Hughes Hallett; John Lewis
Abstract: Empirical fiscal policy reaction functions based on ex-post data cannot be said to describe fiscal policymakers intentions because they utilise data which did not exist when their decisions were made. A characterisation of what fiscal policy makers were trying to do requires real time data. This paper compares fiscal policy reaction functions for 14 European countries over the period 1995-2006 using both types of data. We exploit the information contained in real time and ex-post data and develop a new approach to estimating the automatic and discretionary fiscal policy responses to changing economic conditions. This avoids the uncertainties and inaccuracies associated with filtering the data in an attempt to estimate potential output or the structural budget. We find that the often commented on pro-cyclicality of discretionary policy only arises in the ex-post data; the real time data suggests that policymakers have tried to run counter-cyclical discretionary policy, but find it hard to do so due to data constraints. Compared to elsewhere in the literature, our model yields lower estimates of the automatic fiscal responses and stronger estimates of the discretionary responses to the output gap.
Keywords: discretion; fiscal policy; real time data
JEL Codes: E61; E62
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Fiscal policymakers intend to run countercyclical discretionary policies (E62) | Significant positive coefficient on the output gap (E23) |
| Weak procyclicality in discretionary fiscal policy (E62) | Apparent procyclical nature of fiscal policy in previous studies (E62) |
| Data revisions and measurement errors (C82) | Apparent procyclical nature of fiscal policy in previous studies (E62) |
| Automatic fiscal response is weaker than previously thought (E62) | Discretionary response is stronger (H56) |