Top-down vs. bottom-up? Reconciling the effects of tax and transfer shocks on output

Using the bottom-up approach of Romer and Romer (2010), we construct a rich narrative dataset of net-revenue fiscal shocks for Germany by reconstructing and extending the tax shock series of Hayo and Uhl (2014) and coding a shock series for social security contributions, benefits and transfers. Based on quarterly data for 1974q1 to 2013q4 we estimate the multiplier effects of shocks to net-revenues, taxes, social security contributions and benefits in a proxy SVAR framework (Mertens and Ravn 2013) and compare them with estimates of the top-down identification inspired by Blanchard and Perotti (2002). We find multiplier effects of net-revenue components for Germany between 0 and 1 for both the top-down and bottom-up approaches. These estimates are on the lower end of the scale given in the literature and we discuss the differences.