Government and the financing of health care.

The contribution of the present paper is to provide better estimates of employer-provided health insurance premiums and therefore tax expenditures and, more importantly, to show the distributional effects of each of these programs.

Abstract

Over the past two decades, government financing of health care has increased fourteenfold. In 1960, prior to the enactment of the Medicaid and Medicare programs, public on health care were $6.6 billion; in 1965, they were $11 billion, in 1975 $56 billion, and by 1979 they had reached $91.4 billion. In addition, there was a notable shift from state and local to federal outlays. In 1960, state and local were slightly larger than federal expenditures; by 1975, federal were twice the size of state/local expenditures. The focus of this paper is twofold: I will examine both Medicare and Medicaid and an indirect health care subsidy, namely tax expenditures. This expenditure consists of losses in federal revenue resulting from tax savings for certain groups of taxpayers allowed by the tax code. The logic of calling these revenue losses expenditures is that a reduction in tax liability is equivalent to an expenditure in the same way that a forgiveness of debt is equivalent to a payment (Congressional Budget Office). In the case of health, losses in federal revenue (and thus savings to tax payers) result from the exclusion of employer-paid health insurance premiums from their employees' taxable income, and from the deductibility of certain medical expenses for the purpose of computing personal income tax liabilities. Regarding such tax on health as a major federal program is not a new concept, nor is it unique to compare tax on health with the amounts spent on Medicare and Medicaid (see, for example, Martin Feldstein and Elizabeth Allison; Eugene Steuerle and Ronald Hoffman; Ronald Vogel). The contribution of the present paper is to provide better estimates of employer-provided health insurance premiums and therefore tax and, more importantly, to show the distributional effects of each of these programs.