The economic incidence of social security contributions: A discontinuity approach with linked employer-employee data

We estimate economic incidence of social security contributions (SSC) on the basis of cross-sectional earnings distributions. The approach exploits discontinuities in earnings distributions at kinks in the budget set which are informative about tax incidence. Contrary to most research on SSC incidence, it does not rely on policy reforms, panel data, or hours information. When the location of kinks does not change significantly, estimates represent equilibrium incidence and are less affected by short-run adjustment frictions than results based on policy reforms. We refine the framework proposed by Alvaredo and Saez (2007), discuss identifying assumptions and related problems in empirical applications. We also suggest parametric and non- parametric estimators. The approach is applied to earnings caps of SSC in Germany where the marginal SSC rate drops to zero. The linked employer-employee data used provide precise measures of gross and net earnings. Utilizing two separate earnings distributions improves identification in the presence of measurement error. We find substantial negative discontinuities at most earnings caps of SSC in the distribution of observed net earnings. Together with smooth gross earnings distributions around the caps this provides consistent empirical evidence that legal and economic incidence of SSC coincide.