Abstract
<jats:p>Discrete policy thresholds are pervasive in tax and regulatory systems and can substantially distort behavior. We show that notches acting as barriers to mobility within a distribution generate distortions extending far beyond the threshold. The same mechanism biases conventional difference-in-differences estimators, and we propose a new methodology to recover causal effects. Applying the method to the abolition of a size-based payroll tax notch, we find that the notch reduced the number of firms above the threshold by 18 percent and lowered treated firms' employment, capital stock, and value added by 10 percent, whereas conventional difference-in-differences estimates imply negligible effects.</jats:p>
Show More
Keywords
threshold
conventional
differenceindifferences
effects
notch