Government spending effects on the business cycle in times of crisis

Berger, T.; Dubbert, T.

Abstract

The literature on fiscal multipliers has long established a positive impact of public spending on output. However, the size of this effect strongly depends on the employed identification strategy. Moreover, fiscal multipliers are uninformative as regards the state of the economy. Using counterfactual scenario analyses based on a conditional forecast algorithm in combination with the Beveridge-Nelson decomposition, we address both issues by assessing the effectiveness of public spending in terms of its influence on the output gap. Our approach is independent of the chosen identification strategy and allows us to make (quantitative) statements about potential downsides from public spending measures by looking at its effects on the business cycle. Using a US dataset and analyzing hypothetical government spending scenarios in times of historical crises, we find that, to avoid an overheating of the economy in combination with high inflation and public debt, the dosage of fiscal stimulus is crucial for targeted fiscal policy measures and depends on the severity of the crisis.

Keywords
Fiscal policy; output gap; conditional forecast; scenario analysis; Bayesian vector autoregression