Recent tax reform discussions have centered around two major areas of concern: first, the income base subject to the income tax, and, second, the rate at which income should be taxed.' As an alternative to the present income base definition, some proponents of reform have suggested a broader base that closely approximates the Haig-Simons definition of income.2 This approach involves the inclusion in the base of some presently excluded income items such as interest on state and local bonds, gifts, bequests, devises and inheritances, and so forth; the recognition of appreciation and depreciation in the values of property as income subject to tax; and, finally, the addition to the base of items of imputed income attributable to property ownership or service-yielding assets (for example, imputed rent on owneroccupied houses). With respect to the rate structure, however, little agreement exists among tax reform advocates on the proper rate structure or how the present rate structure might be modified. Galvin, Ture, and Stockfisch3 have suggested the substitution of a flat (proportionate) income tax rate for the existing progressive rate structure. A flat rate would eliminate distortion caused by progression, eliminate the need for relief measures for those taxpayers subject to higher rates, eliminate the bunching problem, and lead to simplification of the tax code. These arguments and the fact that rate progression cannot be justified on a sound theoretical basis make a flat rate structure appealing. However, such a proposal must also be evaluated in terms of the differential incidence of this tax system as opposed to the present one-that is, how many gain and how many lose by the substitution, and how the gainers and losers are distributed by income class. The purpose of this paper is to present empirical evidence on the distribution of