Estimating Dynamic Income Responses to Tax Changes: Evidence from Germany

Abstract

This paper provides new empirical insights on the elasticity of taxable income to the net-oftax rate. Using a panel of German income tax return data, we followed taxpayers from 2001 to 2006 to analyze the effects of the German tax reforms of 2004 and 2005. Implementing a dynamic model as proposed by Holmlund and Söderström (2011), we are able to disentangle short-term and long-term responsiveness. These estimates allow us to distinguish between different dimensions of behavioral changes: short-term income reactions in contrast to ‘real’ changes in (reporting) behavior. We compare our results with recent German estimates from the established approach by Gruber and Saez (2002) applied by Gottfried and Witzcak (2009). Following Chetty`s (2009) theoretical considerations, we use multiple (tax code related) income concepts and alternative sample choices. We provide several robustness and validity analyses of the most common income concept, i.e. taxable income excluding capital. Our preferred specification yields (very) high short-term yet small long-term elasticites. The latter range from 0 to 0.16, implying none or only modest persistent behavioral changes to marginal tax rate cuts.