We estimate taxable income responses using panel data from German adminis-
trative tax records by exploiting several tax reforms between 2001 and 2008. We
provide panel regression evidence as well as graphical evidence. Our results are
sensitive with regard to the chosen estimation technique, income controls and
sample selection. Non-tax related factors (mean reversion and heterogeneous
income trends) seem to play a major role. To shed light on the influence of these
factors, we conduct a Monte Carlo exercise with simulated data to compare the
efficiency of each proposed instrument along with different income controls. We
illustrate for each specification how the estimates are influenced depending on
the income growth process.