The removal of an intertemporal tax discontinuity in Germany provides us with a natural experiment to study the causal effect of taxes on individual stock-trading behavior and the disposition effect. Using individual investor transactions data, we find substantial spikes in selling probabilities around the intertemporal tax discontinuity and no such spikes after the abolishment. In all pre-reform years, the disposition effect is reversed (unusual large) to the left (right) of the tax discontinuity. Using bunching methods, non-parametric regressions, and changes in effective tax rates, we find that investors reduce their effective tax rate by 11.3% due to behavioral responses.

(Florian Buhlmann, Philipp Doerrenberg, Benjamin Loos & Johannes Voget)