Reconsidering the Federal Deductibility of State and Local Taxes
James R. Hines, Jr., Department of Economics, University of Michigan
a 2006 CLOSUP Small Grants Program award
This project will evaluate the economic consequences of the state and local tax deduction, together with the (closely associated) federal exclusion of interest earned on state and local bonds. This issue is long due for reconsideration, since there is considerable reason to believe that there are serious flaws in the received economic logic suggesting that there are large distorations associated with the state and local income tax deduction.
Americans are currently entitled to deduct state and local tax payments in calculating their federal taxable incomes. Deductibility can significantly reduce the real cost of high state and local taxes. For example, a high-income taxpayer facing a marginal federal income tax rate of 35% and a state and local tax rate of 10% in fact has a net state and local tax burden of 6.5%, since federal taxes are reduced by 3.5% as a consequence of federal tax deductibility. Not all taxpayers benefit from the deductibility of state and local taxes, since the benefit requires that that taxpayers itemize deductions, and many find the standard deduction alternative to be more attractive. High-income taxpayers, and those in high tax states, tend to be the ones who benefit from deductibility. In a typical year, 38% of the Michigan taxpaying population itemizes deductions and thereby claims a deduction for state tax payments; nationally, the figure is closer to 35%.
The benefit of tax deductibility is considerable, due in part to the fact that the minority of the population that itemizes its deductions nonetheless pays the vast bulk of U.S. income taxes. The Joint Committee on Taxation estimates that the deductibility of state and local income taxes reduced federal tax revenues by $55 billion in fiscal year 2005. This, together with the estimated $27 billion saving from the exclusion of interest on state and local bonds, produces a total saving of $82 billion, or 5.5% of the aggregate budgets of state and local governments in 2005. Some observers have expressed concern over the magnitude of the implied federal subsidy for state and local spending, and its possible distortionary effect in encouraging subnational governments to undertake projects a portion of which the federal treasury unwittingly underwrites. For example, the 2005 Presidents Advisory Panel on Federal Tax Reforms offered two recommendations for fundamental reform of the nations tax system, both of which included removal of the deductibility of state and local income taxes. The Panel was motivated by the need to raise tax revenue in order to finance other tax reductions, and the preceived distortion associated with the implied federal subsidy to state and local operations.
The purpose of the proposed research project is to evaluate the economic consequences of the state and local tax deduction, together with the (closely associated) federal exclusion of interest earned on state and local bonds. This issue is long due for reconsideration, since there is considerable reason to believe that there are serious flaws in the received economic logic suggesting that there are large distortions associated with the state and local income tax deduction.
The standard economic argument concerns incentives. If the marginal dollar used to finance state and local activities comes from taxpayers facing an average federal income tax rate of 25%, then the U.S. government effectively pays 25% of the cost of any additional spending. Hence a state project that costs $100 and produces only $90 of value would nonetheless look attractive to state policy makers, since the real net cost to state residents is only $75. The economics literature occasionally makes reference to luxury trash removal services. These luxury services are net losers from the standpoint of the country as a whole, since they cost $100 and produce value of just $90, but are undertaken because state and local governments do not bear the full costs of their actions.
Probing these issues just a bit more deeply starts to reveal the weaknesses of the standard argument. Suppose that, instead of trash removal services, the state and local spending category is primary education, and that the return to education comes much later in the form of higher wages and better citizenship. Higher wages are subject to federal taxation. Hence a state government that spends $100 to increase the productivity of its workers by $120 in present value in fact increases their after-tax wages by only $90 if the workers are subject to a 25% federal income tax. If state and local taxes are not deductible, then the state government, acting on behalf of its citizens, will not have incentives to spend this $100, despite the $120 of value that it creates for the country. (There is an additional issue, also noteworthy, that states might spend money to educate workers who will ultimately move to other states and pay taxes there as adults, all of which suggests that state education spending should be subsidized rather than discouraged by the federal government. This consideration will be addressed in the research, but for now we confine attention to cases in which children can be expected to live in the same states throughout their lives.) What is essential to the education example is that the return to state spending comes primarily in a form (higher wages) that is subject to federal taxation. Education spending is by no means the only such example. State and local spending on roads and highways serves two functions, the first to facilitate commerce (including business travel, workers commuting to their jobs, trucks transporting products, and other uses), and the second to improve the quality of personal life by permitting people to get from here to there. The first of these uses, which is likely to be the most quantitatively important, produces returns subject to federal tax. It is noteworthy that even the second of these uses may have some federal tax implications, since the provision of convenient roads that individuals use to shop for food or other items encourages earning income by increasing the net returns to having income to spend.
Viewed in an appropriate light, much of state and local government expenditure is devoted to improving the livelihoods, and therefore the taxable incomes, of local residents. This is true of police and fire services, health and hospital expenditures, and many others. It follows, therefore, that in the absence of federal tax deductibility, the overhang of the federal income tax system would inefficiently discourage state and local governments from spending money. If the return to most state and local spending comes in the form of increasing taxable incomes, then the federal deductibility of state and local taxes serves to undo this distortion, and, by implication, removal of deductibility would introduce important inefficiencies by discouraging spending. Surprisingly, this elementary insight appears nowhere in the economics literature on federal tax deductibility.
The research project on state and local tax deductibility has four components. The first component is to develop the theory of efficient state and local spending levels under federal income taxation, and identify the implications of this consideration for previous research. The second component of the research is to analyze the important state and local spending categories in an effort to measure directly the extent to which state and local spending generates returns that are taxed by the federal government. This portion of the research will be based largely on the analysis of others, aggregated by spending category.
The third component of the research will be an econometric investigation of the determinants of state and local spending levels. Previous studies estimate state and local spending levels as functions of the fraction of state income taxes that are deductible on federal returns, interacted with marginal federal income tax rates. The implication of the pecuniary view of the returns to state and local spending is that federal income tax rates should not matter, if they are expected to remain constant; in this interpretation, only deductibility matters. In a somewhat more sophisticated framework, what should matter is deductibility and the expected change in federal marginal tax rates between the time when expenditures are made and the time when returns are received. Furthermore, incentives vary by spending category, depending on the length of time before returns are realized. These propositions are testable, and this portion of the research will test whether actual spending patterns correspond to these predictions. Expenditure data are available from the Census of Governments, and tax data from the World Tax Database and the NBER TaxSim program.
The fourth component of the research is to investigate the extent to which the federal tax exclusion for state and local bond interest may subsidize state and local spending, and thereby indirectly substitute for federal income tax deductibility in the case the latter is removed. Clearly this tax exclusion is valuable, though state and local governments can obtain some if not all of the same benefits by taxing themselves in early years, thereby avoiding the need for taxpayers to hold assets whose returns are taxed by the federal government. Given the extensive use of debt finance by state and local governments, and the correspondingly sizable magnitude of the federal tax subsidy, it is important to determine the role that the tax subsidy for debt finance may play in attenuating the effects of federal taxation of the returns to state and local spending.