Current Research Projects

Combatting Tax Evasion: Evidence from Comparing Commercial and Business Tax Registry


In 2008 and 2014, the South African Revenue Service (SARS) did snapshot synchronizations of its business tax registry with the country's commercial register in an attempt to identify firms that are non-compliant with their obligation to register with SARS for business tax purposes. We analyse these interventions drawing on SARS's business tax registry and the population of business tax returns between 2009 and 2014. Several findings emerge. First, in both years, the comparisons resulted in the identification of around 300,000 non-compliant taxpayers, providing prima facie evidence of significant extensive-margin tax evasion. The interventions significantly raised South African business tax revenues in the following years despite the fact that the identified `extensive-margin evaders' exhibit a lower propensity to submit tax returns and, conditional on return submission, report less income than comparable entities that voluntarily registered with SARS. The analysis further suggests that the observed gap in reported taxable income relates to underlying differences in firm size and corporate productivity rather than intensive-margin tax evasion. In line with `missing middle theories', extensive-margin evaders that submit tax returns are, moreover, found to exhibit increased sales and asset growth after their forced registration with SARS.


A Data-Driven Procedure to Determine the Bunching Window
An Application to the Netherlands


We extend the bunching approach introduced by Saez (2010) by proposing an intuitive, data-driven procedure to determine the bunching window. Our method explicitly allows the bunching window to be asymmetric around the threshold and to be more flexible. It also enhances the reproducibility of studies implementing the bunching approach. In our application to identify taxpayers' responsiveness to taxation for the Netherlands, we find clear evidence of bunching behaviour at all three thresholds of the Dutch tax schedule with a precise estimated elasticity of 0.023 at the upper threshold. We find much larger estimates for women and self-employed individuals. We also identify significant bunching behaviour for individuals in paid employment, which we can mostly attribute to tax deductions that can be shifted between married tax filers. Since bunching is absent among single tax filers, we conclude that real responses to taxation are modest.