Discrete earnings responses to tax incentives: Empirical evidence and implications

We study the consequences of discrete rather than continuous earnings choice sets on individual responses to income taxes. In our empirical application, we utilize an income notch created by the study subsidy system for higher education students in Finland and a reform that shifted out the location of this notch. We find that the reform, which changed the income tax schedule by increasing the location of the notch from 9,000 euros to 12,000 euros, affected the income distribution from earnings of about 2,000 euros onward. Because the tax schedule did not change around these lower incomes, the wide-ranging response to the reform constitutes a puzzle from a standard model point of view. We develop further results, theoretical arguments and a simulation model that all suggest that the shifting of the distribution can be explained with discrete earnings choices. Moreover, we discuss the welfare implications of discrete earnings choices, and find that welfare losses can be greater than empirically estimated if the underlying behavior is constrained by discrete earnings and they are thought to be represented by continuous earnings choices.