Fiscal Policy and Economic Activity
Conventionally, macroeconomic consequences of economic policy have been researched in structural vector autoregressions (SVAR). Fiscal policy SVAR models estimated for the U.S. are usually supportive for large fiscal policy effects on output. In one prominent example, the tax multiplier is estimated...
|Online Access:||PDF Full Text|
No Tags, Be the first to tag this record!
|Summary:||Conventionally, macroeconomic consequences of economic policy have been researched in structural vector autoregressions (SVAR). Fiscal policy SVAR models estimated for the U.S. are usually supportive for large fiscal policy effects on output. In one prominent example, the tax multiplier is estimated as -1.33, and the spending multiplier as 1.29 (Blanchard and Perotti, 2002). SVAR models require untestable identification assumptions; thus, prompting the search for natural experiments as an alternative source of identification. In a particularly influential study, Romer and Romer (2010) construct a historical account of exogenous legislated U.S. tax changes and estimate a much larger tax multiplier of around -3. Applications of the SVAR methodology to Germany have generally found rather small effects of fiscal policy on output. In Hayo and Uhl (2014a), we use a natural experiment approach, closely following Romer and Romer (2010), and find strong effects of tax changes on output. Based on our evidence, one can conjecture that the tax multiplier in Germany might be as large as -2.4. The estimated tax multipliers are much larger than alternative estimates derived in fiscal policy VAR models for Germany. Implementing this study required intensive data collection processes; Uhl (2013) contains the documentation of these efforts.
Most studies on the macroeconomic consequences of fiscal policy use aggregate nationwide data. In Hayo and Uhl (2014b), we estimate the consequences of federal tax policy actions for regional economic activity in the U.S. We find considerable variation in how regional output reacts to federal tax changes and that estimated state multipliers range between
–0.2 in Utah and –3.7 in Hawaii. An econometric analysis of determinants behind these differences reveals that the size and composition of a state tax base is related to the strength of the local income reaction. These results improve our understanding of the precise transmission mechanism of fiscal policy shocks. In Uhl (2014), I estimate the consequences of U.S. state-level fiscal policies for local economic activity and conclude that state-level spending multipliers are relatively small, while tax multipliers are large. These results allow for assessing the consequences of subnational fiscal policies and provide stylized-facts on fiscal multipliers in a monetary union. It is interesting to note that estimated multipliers at the state level are comparable to estimates derived at the country level despite their different transmission mechanism. I also find that both increases in state spending and in state taxes improve out-of-state output which suggests that spillovers among states or countries are relevant.
Inference on ‘fiscal multipliers’ in aggregate time series requires untestable identification assumptions. Asking economic agents directly about their responses to fiscal policy is an appealing non-standard alternative. Shapiro and Slemrod (1995), and follow-up papers, ask U.S. residents about their consumption responses to various tax changes. We extend on this research by directly asking the German population about their consumption and labor supply responses to a recent 2013 payroll tax change using a representative population survey (Hayo et al., 2014, Hayo and Uhl, 2014c, and Hayo and Uhl, 2014d). About 55 per cent of the respondents indicate that they have increased spending; suggesting that tax changes in Germany have a relatively large impact on consumption and, hence, on economic activity. Based on the evidence from this representative survey, the effects of tax changes on labor supply, however, are likely small. The relative dominance of consumption responses, vis-à-vis labor supply responses, is a conclusion that is also present in the aggregate time series evidence in Hayo and Uhl (2014a). One further noteworthy implication from our representative survey is that currently low interest rates reduce incentives to save as well as incentives for labor supply.|