The Marginal Welfare Cost of Income Redistribution

Abstract

Arthur Okun [14] has recently emphasized the importance of evaluating the leakages that occur when the government redistributes income from one group to another. A number of studies have, of course, examined the leakages, or welfare costs, that result from certain taxes and transfer programs, but the measure of welfare cost appropriate for evaluating the amount of redistribution has neither been clarified nor estimated in the existing literature. Clearly the relevant concept is some measure of marginal welfare cost that results from a change in the amount of redistribution, and studies which estimate the total welfare cost of specified programs provide (as I will show) little clue to its magnitude. In this paper I develop estimates of two different measures of the marginal welfare cost of redistribution for several income transfer programs. Changes in the volume of redistribution are assumed to be effected through a negative income tax (NIT) coupled with positive income taxation to supply the necessary revenues, and with constant marginal tax rates used in both the tax and transfer policies. Analytically, this appears to be the simplest form of redistributive policy, and it also enjoys strong support in the economics profession. Part of the support for this type of policy seems to derive from a belief that broad-based taxes and transfers are likely to be less distorting than alternative policy options. While many types of distortions are produced by income transfers and taxes, perhaps the most important (judging from the attention given it) is the distortion of labor supply decisions, and this distortion is the focus of the present paper. There is empirical evidence suggesting that the elasticity of labor supply tends to be relatively low, and this would seem to imply that welfare costs would also be relatively low [17; 10]. However, despite the assumption of low values for labor supply elasticities, my