David Gamage (UC-Berkeley) presents On Double-Distortion Arguments, Distribution Policy, and the Optimal Tax Mix at UC-Berkeley today as part of its Economics Department Public Finance Workshop Series:
How should we tax the wealthy? Should non-tax legal rules promote distributional equity, or focus solely on efficiency? An influential position in the law and economics and tax policy literatures concludes that—with some limited exceptions —only a federal tax on labor income should be used for distribution. Variations on this position are sometimes called “double-distortion” arguments.
A double-distortion argument underlies Kaplow and Shavell’s influential claim that non-tax legal rules should be designed solely based on efficiency, leaving all distributional considerations to the tax system. Within the tax literature, double-distortion arguments have led many scholars to conclude that we should only tax either labor income or consumption -- that we should not tax wealth, capital income, or bequests. Following similar logic, with respect to state and local governments, the dominant models of fiscal federalism imply that subnational governments should leave most distributive concerns to the federal government. In these, and a variety of other policy contexts, double distortion models have been used to argue that nearly all distribution policy should be conducted through a federal tax on labor income, such that most other tax instruments and non-tax legal rules should be designed solely to promote efficiency.
This Article begins with the observation that double-distortion models typically assume the existence of a near-perfect labor-income tax. Specifically, most double-distortion models assume that a labor-income tax can accurately measure taxpayers’ ability to earn labor income with only a single limitation: that taxpayers can shift from labor to leisure in order to replace taxed monetary consumption with untaxed leisure consumption and thereby reduce their labor-income-tax liabilities. Very few double-distortion models account for the possibility of taxpayers reducing their labor-income-tax liabilities by engaging in tax avoidance, tax evasion, or other tax-reduction techniques other than by shifting from labor to leisure.
In contrast to the assumption that the possibility of labor-to-leisure distortions is the sole flaw of the labor-income tax, the empirical literature suggests that taxpayers engage in wide variety of diverse techniques for reducing their labor-income-tax liabilities. This appears to especially be true with respect to high-income taxpayers, for whom distributional considerations are particularly relevant. Indeed, there is essentially no evidence that high-income taxpayers significantly reduce their labor effort in response to taxation. In contrast, there is a plethora of evidence documenting that high-income taxpayers engage in a variety of other tax minimization strategies.
This Article analyzes how double-distortion models can be generalized to account for the possibility of tax-reduction techniques other than labor-to-leisure distortions. Under plausible assumptions, this Article argues that generalizing double-distortion models in this fashion yields significantly different policy implications.