Tax policy.

Metcalf, Gilbert E.

My research on taxation has focused on three areas: taxes and government financing in a federal system; measuring tax incidence; and energy-related taxation. What is striking in many of the projects that I describe here is the range of effects that taxes can have on behavior, many of which are surprising.

Financing Government in a Federal System

Taxes at the national level interact with the financing of state and local governments in a variety of ways. Many state and local taxes are deductible at the federal level, thereby reducing the cost of raising a dollar of state or local tax revenue for federal itemizers. Income from municipal debt in large part is untaxed at the national level. Equally important, income earned by the state and local sector is not subject to taxation. This provides a variety of arbitrage opportunities that can help explain the financial behavior of state and local governments.

Deductibility of state and local taxes became an important issue during the debate leading up to the Tax Reform Act of 1986 (TRA86). The elimination of this deduction was a major source of revenue to pay for lower marginal tax rates in the Treasury I plan. However, my research with Martin Feldstein called into question the revenue response predicted in Treasury 1.(1) We argued that eliminating the personal deduction for state and local taxes would lead to a shift away from these taxes toward others that continue to be deductible at the corporate and business level. Using the NBER TAXSIM model to construct state-level average tax prices for state and local taxes, we found large price elasticities for tax shares.

One implication of this finding is that eliminating deductibility for state and local taxes would not necessarily increase federal tax receipts. Interestingly, we also found that changes in deductibility have large effects on the share of different tax instruments, but little effect on the overall level of spending at the state and local level. While there is still some controversy over the effect of changing tax prices on the overall level of state and local spending, subsequent research has found substantial price effects for tax shares in a variety of models.(2)

Ultimately, TRA86 eliminated deductibility of only state and local general sales taxes. Based on the research just cited, economists predicted that the use of general sales taxes would decline. Instead, it appears that state and local governments relied more heavily on sales taxes after 1986 than before. I rationalize this behavior with the strong price effects found in the previous literature,(3) because TRA86 increased the tax price for all state and local taxes, not just for the sales tax. For an itemizer at the federal level, the tax price of a deductible tax is one minus the federal marginal tax rate. Decreases in marginal tax rates at the federal level thus increase the tax price for deductible taxes. While it might appear that the tax price for sales taxes would increase by more than the tax price for deductible taxes, this will not be true if there is not complete deductibility of sales taxes at the margin prior to 1986. Evidence suggests that the sales tax "look-up" tables used by most taxpayers to compute their deduction underestimated at the margin the sales tax liability actually incurred.(4)

TRA86 provided a natural experiment for studying the incentive effects of federal deductibility on the state and local tax structure and on state and local spending. Using data on state governments over a nine-year period and the Bureau's TAXSIM model, I find that state income taxes are sensitive to changes in their tax price.(5) There is also evidence in the data that high-income groups are more concerned about income taxes and middle-income groups are more concerned about sales taxes.

In addition to affecting the choice of tax instruments, federal taxation affects saving and borrowing decisions by state and local governments. Roger H. Gordon and Joel B. Slemrod have noted several arbitrage opportunities available to state and local governments because of the failure to tax their earned income and the failure to tax income on bonds they issued.(6)

In a series of papers, I have used panel datasets on state and local governments to study these different arbitrage opportunities. The simplest arbitrage opportunity available to state and local governments is to borrow funds at the tax-exempt rate and invest them in taxable (higher-yielding) securities. Although federal regulation attempts to prohibit this activity, I find strong evidence that this form of arbitrage was occurring prior to TRA86.(7) Arbitrage possibilities can be used to build a model of the supply of municipal bonds. In effect, communities must decide how to raise funds; they can do so by borrowing publicly (issuing municipal debt) or by borrowing privately (taxing). This suggests that the spread between the aftertax rate of return available to residents in a community and the municipal borrowing rate for bonds issued by that community affects the supply of municipal debt. I find strong support for this model using data on both state and local governments.(8)

In another paper, Gordon and I explore the implications of these arbitrage opportunities for the size of the subsidy for municipal capital arising from tax-exempt municipal debt.(9) We argue, that the exemption provides little in the way of subsidies to municipal capital, but rather provides subsidies to individuals in high and low tax brackets. it subsidizes high-income people to the extent that tax-exempt interest rates are lower than the aftertax rates of return they can earn on equivalent taxable securities. It subsidizes low-income people by allowing them to borrow funds at lower rates of interest (through their community) than through private markets. In addition, the tax revenue that the federal treasury would obtain by eliminating this tax exemption would be small, given the decline in municipal borrowing that would occur if the tax exemption were eliminated.

Tax Incidence

So far I have considered how governments choose to allocate their revenue requirements among different tax and borrowing instruments. Once that decision is made, it is important to understand its distributional effects. There has been considerable interest recently in measuring the incidence of taxes over the lifetime of taxpayers.(10) I find that the system of state and local taxes is significantly more progressive over individuals' lifetimes than over a year.(11) In fact, using data from the Consumer Expenditure Survey, I find that general sales taxes are progressive over the life cycle, and as progressive as state and local income taxes. Moving from annual to lifetime measure shifts sales taxes toward proportionality. Exempting necessities, as most states do, then moves the tax to progressivity.

If we ignore bequests, the lifetime budget constraint suggests that a flat tax on consumption should be proportional over the lifetime. Since bequests are small for the bulk of the population, this suggests that a value-added tax (VAT) with exemptions for necessities easily could be progressive over the lifetime. This is borne out in research that Erik Caspersen and I are conducting using data from the Consumer Expenditure Survey and the Panel Study of income Dynamics.(12) Using two different measures of lifetime income, we find that a VAT that excludes housing, health, and food costs is moderately progressive over the lifetime. This contrasts sharply from the results using an annual framework, in which the tax appears distinctly regressive.

Taxes and Investments in Energy Efficiency

Do people respond to tax incentives for energy conservation investment by increasing their rate of investment in energy-efficient capital? While one might expect a clear-cut affirmative response, in fact there has been little evidence to support such a claim. Studies of tax incentives for conservation investment typically have used variation in state programs to identify the effects of tax incentives. Kevin A. Hassett and I argue that a major reason for the lack of a response is the endogeneity of state tax incentive programs.(13) If residents of a particular state are inclined to make investments in energy-efficient capital, then the states may feel less need to create a tax incentive to induce further investment. Using data from a panel of tax returns, we are able to construct a consistent measure of the effect of tax incentives on the probability to invest in conservation capital. We find this effect to be puzzlingly small. We argue that this is probably because these investments are both irreversible and risky. We then apply the models of Dixit and Pindyck to explain investment in energy-efficient capital.(14) This helps to explain the findings by Hausman and others of very high discount rates for investment in energy-efficient capital.(15)

(1) M. Feldstein and G. E. Metcalf, "The Effect of Federal Tax Deductibility on State and Local Taxes and Spending," NBER Reprint No. 911, September 1987, and Journal of Political Economy 95 (1987), pp. 710-736. (2) D. Holtz-Eakin and H. S. Rosen, "Tax Deductibility and Municipal Budget Structure," NBER Working Paper No. 2224, April 1987; L. B. Lindsey, "Federal Deductibility of State and Local Taxes: A Test of Public Choice by Representative Government," NBER Working Paper No. 2292, June 1987; and G. Zodrow, "Eliminating State and Local Tax Deductibility: A General Equilibrium Model of Revenue of Effects"; and all three papers in Fiscal Federalism: Quantitative Studies, H. S. Rosen, ed. Chicago: University of Chicago Press, 1988. See also R. P. Inman, "The Local Decision to Tax: Evidence from Large U.S. Cities," NBER Reprint No. 1406, May 1990, and Regional Science and Urban Economics 19 (1989), pp. 455-491. (3) G. E. Metacalf, "Deductibility and Optimal State and Local Fiscal Policy," Economics Letters 39 (1992), pp. 217-221. (4) R. Ebel, "Comment on |Tax Exporting, Federal Deductibility, and State Tax Structure,'" Journal of Policy Analysis and Management 12 (1993), pp. 127-130. (5) G. E. Metcalf, "Tax Exporting, Federal Deductibility, and State Tax Structure," Journal of Policy Analysis and Management 12 (1993), pp. 109-126. (6) R. H. Gordon and J. B. Slemrod, "An Empirical Examination of Municipal Financial Policy," in Studies in State and Local Public Finance, H. S. Rosen, ed. Chicago: University of Chicago Press, 1986. (7) G. E. Metcalf, "Arbitrage and the Savings Behavior of State Governments," Review of Economics and Statistics 72 (1990), pp. 390-396. (8) G. E. Metcalf, "Federal Taxation and the Supply of State Debt," Journal of Public Economics, forthcoming, and "The Role of Federal Taxation in the Supply of Municipal Bonds: Evidence from Municipal Governments," National Tax Journal 44 (1991), pp. 57-70. (9) R. H. Gordon and G. E. Metcalf, "Do Tax-Exempt Bonds Really Subsidize Municipal Capital?" NBER Working Paper No. 3835, September 1991, and National Tax Journal 44 (1991), pp. 71-80. (10) D. Fullerton and D. L. Rogers, "Lifetime Versus Annual Perspectives on Tax Incidence," NBER Working Paper No. 3750, June 1991, and J. M. Poterba, "Is the Gasoline Tax Regressive?" NBER Working Paper No. 3578, January 1991, and Tax Policy and the Economy 5 (1991), pp. 145-164. (11) G. E. Metcalf, "The Lifetime Incidence of State and Local Taxes: Measuring Changes During the 1980s," NBER Working Paper No. 4252, January 1993. (12) E. Caspersen and G. E. Metcalf, "Is a Value-Added Tax Progressive? Annual Versus Lifetime Incidence Measures," mimeo, Department of Economics, Princeton University, February 1993. (13) K. A. Hassett and G. E. Metcalf, "Energy Tax Credits and Residential Conservation Investment," NBER Working Paper No. 4020, March 1992. (14) A. K. Dixit, "Investment and Hysteresis," Journal of Economic Perspectives 6 (1992), pp. 107-132, and R. S. Pindyck, "Irreversible Investment, Capacity Choice, and the Value of the Firm," NBER Working Paper No. 1980, July 1986, and American Economic Review 78 (1988), pp. 969-985. (15) J. A. Hausman, "Individual Discount Rates and the Purchase and Utilization of Energy-Using Durable," Bell Journal of Economics 10 (1979), pp. 33-54.