How Do Taxes Affect People´s Behavior?

Most advanced economies of the world rely heavily on the taxation of labor income to finance public spending with, on average, one-third of all government revenues coming from this source of taxation (see OECD 2013). However, taxing labor income to this extent does not come without costs, raising questions of adverse incentives to work for the taxed individuals, for example. To find the optimal and most efficient design of income taxation is a frequent challenge for policy-makers around the world. Hence, measuring behavioral responses to taxation occupies a critical place in public economics. In principle, economic literature distinguishes between real and avoidance/evasion responses to taxation (Slemrod 1995). This research project aims to shed new light on both of them. First, in recent years a vivid literature emerged using large administrative data and compelling identification strategies to estimate real responses to income taxation. A promising new avenue was laid out by Saez (2010), proposing a bunching estimator to uncover the elasticity of labor supply. His approach is based on the intuition that discontinuous reductions in wage rates at kink points of a tax schedule yield a non-parametric source of identification for the underlying elasticity. His method has been tested only a few times since data requirements are large. As a result, the question whether wage earners respond with bunching to changes in tax rates is unsettled. Furthermore, the identifying assumption of the bunching estimator is that the income distribution would be smooth in the absence of discontinuities in tax rates; an assumption existing studies were unable to test. Making use of a comprehensive tax reform in Austria introducing a large first kink point allows me to test this assumption and establish the first non-parametrically identified evidence of earnings elasticities combining the bunching procedure with a tax reform. The second part of this project turns its focus on evasion responses to taxation, arguably the most difficult response to study. Lenient enforcement of a self-reported commuter tax allowance puts me in a position to observe compliance behavior for a large group of taxpayers quite precisely (see Paetzold and Winner 2014). Specifically, combining several administrative datasets allows me to test for the different channels through which compliance behavior is transmitted, including networks within firms as well as families. Hence, this part of the research project aims to produce new insights about why some people do cheat on their taxes while others pay them honestly, with policy implications spanning from practical tax collecting and enforcement strategies to the optimal design of tax systems.

Most advanced economies of the world rely heavily on the taxation of labor income to finance public spending with, on average, one-third of all government revenues coming from this source of taxation (see OECD 2013). However, taxing labor income to this extent does not come without costs, raising questions of adverse incentives to work for the taxed individuals, for example. To find the optimal and most efficient design of income taxation is a frequent challenge for policy- makers around the world. Hence, measuring behavioral responses to taxation occupies a critical place in public economics. In principle, economic literature distinguishes between real and avoidance/evasion responses to taxation (Slemrod 1995). This research project sheds new light on both of them. The first part of the project uses a large jump in the Austrian income tax schedule where marginal tax rates increase from zero to 38 percent to estimate (real) labor supply responses of wage earners. My study reveals a sharp response of wage earners to this tax rate change. More precisely, I find especially women and secondary earners adjusting their labor supply in order to avoid this first tax bracket. The results suggest that the sharp increase in part-time working can (partly) be attributed to the incentives set by the income tax schedule. Furthermore, I also observe taxpayers to respond along other margins than labor supply, for instance through increasing deductions when crossing the tax bracket. In sum, my findings suggest that a tax system with large jumps in tax rates at the bottom can substantially distort the earnings distribution, and calls for a better and more efficient design of the income tax schedule. The second part of the project focuses on evasion responses to taxation, arguably the most difficult response to study. Specifically, it aims to answer the question whether tax evasion runs in the family. The unique combination of various administrative datasets allows to observe evasion behavior with respect to a widely used commuter tax allowance across two generations. This setting provides a rare opportunity to gain novel insights into how compliance behavior is transmitted from parents to their children. The results suggest that the individual compliance decision is significantly influenced by family behavior, with paternal non-compliance increasing a children's non-compliance by about 25 percent. These findings have important policy implications. First, it shows how evasion opportunities caused by loopholes in the tax law decades ago can lower tax compliance also for much younger cohorts. Second, given that tax evasion behavior is causally linked across generations, policies that are able to reduce tax evasion today will have important spillover effects on the next generation. These insights provide important implications for practical tax collecting as well as tax enforcement strategies.