Work disincentive effects of taxes: An empirical analysis of Swedish men

This study evaluates the sensitivity of estimates of substitution and income effects to a wide variety of economic and empirical assumptions. We implement two maximum-likelihood approaches to measure work disincentive effects in Sweden using cross-section data: the first specifies the tax schedule as a piecewise-linear function; and the second approximates this schedule by differentiable functions. We introduce instrumental-variable procedures that are direct analogues to these approaches to determine the importance of various assumptions. Our analysis leads us to two major conclusions: both maximum-likelihood techniques yield nearly identical empirical results; and the exogeneity assumptions maintained by these procedures are highly questionable.