Brookings Institution Fellow Suggests Ways to Make Tax System Neutral

Tax neutrality is a widely accepted concept in principle. In practice, however, tradeoffs between different concepts of neutrality and different goals can be difficult to resolve. But in several cases this concept can provide a useful way to cut through some of the debates about tax policy and identify a more economically efficient way to organize the tax system.

In my testimony I first provide a general introduction to the concept of neutrality and then applications to five specific areas of tax policy. To preview my substantive conclusions in these areas:

The primary purpose of the tax system is to raise the revenue needed to pay for government spending. As such, the goal is to raise this revenue without distorting the decisions that individuals and firms would otherwise make for purely economic reasons. For example, an efficient economic system people would choose between chocolate chip cookies and oatmeal cookies based on their own personal tastes and the costs of these products. If policymakers imposed a tax on chocolate chip cookies but not on oatmeal cookies the result would be that people would now factor taxes into their choice about which type of cookie to consume -- and possibly end up consuming the less desirable cookie because it was cheaper.

In addition to distorting choices, non-neutralities in the tax system also lead people and firms to devote more socially wasteful effort to transforming the form or substance of their activities to reduce their tax payments, for example by hiring lawyers and accountants to structure financial transactions in a manner that minimizes tax liability.

In some cases deviations from a neutral tax system are unavoidable. It is widely agreed that tax payments should increase with some measure of well-being, like income, consumption or wages. One inevitable consequence of this agreement is that the market consumption of goods and services will be taxed, either directly (as in a consumption tax) or indirectly (as in an income or wage tax, both of which tax the money used to purchase consumption goods). Time spent outside of work, what economists label as "leisure," is not taxed. As a result, people will consume relatively more leisure -- which is equivalent to a reduction in labor supply. Whether this is a quantitatively large or important effect is another question, but at a conceptual level this is a way that the tax system departs from the neutral ideal.

In other cases, deviations from a neutral tax system reflect the goals of policymakers. The tax system is designed to encourage home ownership, contributions to charity, health insurance, and higher education and to discourage smoking and drinking alcohol. Whether these goals are all appropriate or the tax system is the best way to achieve them is another question, some aspects of which will be discussed further below.

Five Applications of Neutrality to Policy Issues

This general discussion motivates the application of the concept of neutrality to five specific issues.

(1) Overall Tax Reform: A Broader Base and Lower Rates

One of the traditional mantras of tax reform is to "broaden the base and lower the rates." This involves two objectives: (1) broadening the base helps make the tax code more neutral between different activities by including more types of income in the definition of income and allowing fewer deductions and credits for specified activities, and (2) lowering tax rates makes the tax code more neutral about the choice between working and not working. Both halves of the process potentially improve efficiency.

One way to gauge the deviation of the tax code from the ideal of a pure income tax is to examine tax expenditures which are defined in statute as "revenue losses attributable to provisions of the Federal tax laws which allow a special exclusion, exemption, or deduction from gross income or which provide a special credit, a preferential rate of tax, or a deferral of liability."² The Office of Management and Budget and the Joint Committee on Taxation regularly release itemized reports on tax expenditures. In the last budget the Treasury listed a total of $987 billion in tax expenditures for FY 2008, including $878 billion for individuals and $108 billion for corporations.³ This total approaches the total amount of discretionary spending ($1,114 billion in FY 2008) and non-interest mandatory spending ($1,527 billion in FY 2008).

If, for example, $500 billion worth of tax expenditures were eliminated that would permit a 32 percent reduction in all individual and corporate income tax rates. Economists generally presume that a tax code with a broader base and lower rates will be more efficient and conducive to economic growth. A number of specific proposals have been made that embody these basic principles, including Senator Ron Wyden and Congressman Rahm Emanuel's Fair Flat Tax Act (S. 1111) and the proposals put forth by the President's Advisory Panel on Federal Tax Reform.

(2) Using the Tax Code to Encourage Desired Behavior: Credits Instead of Deductions

In some cases, policymakers may want to encourage desired activities like homeownership or a college education. In these cases it is worth examining whether the specific goal could be better accomplished through a spending program or through the tax code. In many cases a spending program can be more effectively targeted and delivered to serve the goal in question. But in some cases subsidizing these activities through the tax code may be more efficient. Although administering social programs through the tax code increases the burden on the Internal Revenue Service and can increase the complexity of tax returns. But if these tax expenditures were converted to spending programs, that complexity would simply be shifted to another government agency. Duplicative paperwork would, in fact, likely increase the overall administrative burden for the government, not to mention the burden on families struggling to provide the same information on multiple forms to multiple government agencies. Moreover, many spending programs phase out benefits as incomes rise in a manner that is not fully transparent and not integrated across programs. As a result, it is common for beneficiaries to discover that, for every $1 they earn, they lose 50 cents to $1 in reduced benefits plus higher taxes. Locating social expenditures in the tax code makes their phase-out rates more transparent and easier to harmonize in order to prevent the effective marginal tax rates in excess of 50 or even 100 percent that are often observed in the current tax and transfer system.

But to accept -- and in some cases even embrace -- tax expenditures is not to defend how they are presently structured. For years tax analysts of widely differing philosophies have written about the benefits of shifting tax expenditures from deductions to uniform, refundable tax credits. A deduction of $1 is worth 35 cents to someone in the 35 percent marginal tax bracket, but only 15 cents to someone in the 15 percent bracket. A credit, by contrast, provides the same tax subsidy regardless of one's tax bracket, and a refundable credit does so even if the credit exceeds one's total tax liability. Lily Batchelder, Fred Goldberg, and Peter Orszag laid out the most comprehensive case for the efficiency benefits of using credits rather than deductions to encourage desired behavior that may have broader benefits than accrue to the individual alone (what economists call positive externalities).⁴

Batchelder, Goldberg and Orszag point out that the goal of tax expenditures is often to encourage people to consume more of something -- for example, college attendance. But since deductions reduce the after-tax price more for high-income families than for low-income families, they generally produce too much added consumption by the former and too little by the latter. In the absence of evidence on how much families at different income levels will respond to a given tax incentive, the authors suggest that credits should be the same for all. In reality, however, it is likely to be more economically efficient to make subsidies progressive, with larger subsidies to lower-income households. For example, a uniform credit might be too little to encourage a lower-income family to purchase health insurance, yet more than enough for a high-income household that would have purchased the insurance in any case. If credits are to be effective in encouraging behavior among low-income households, it is also critical that they be refundable.

It is also important to design tax credits so they are non-neutral in the ways that are economically efficient and maximize their cost effectiveness. For example, many tax expenditures are designed to encourage specific activities, whether owning a home or going to college. But policymakers facing limited budgets may not be interested in subsidizing more of these activities -- owning a larger home or going to a more expensive college. In these cases the subsidy should take the form of a flat credit for undertaking the activity, or be capped at a certain level. In this way, the tax code would not be neutral between owning a home and not owning a home, but would be neutral between owning a medium-sized home or owning a large home.

These considerations are especially important in designing the tax treatment of health insurance, the topic I turn to in the next section.

At a minimum, any new tax expenditures with a behavioral motivation should be implemented as credits rather than deductions and should be based on sound behavioral considerations. But the big gains will come only from reforming the existing system of tax expenditures. These reforms could be designed in a manner that also serves other goals, like reducing the nation's large fiscal gap or offsetting some of the increase in inequality in recent decades. But even a revenue-and distribution-neutral reform of tax expenditures would pay substantial dividends, making the tax code fairer and more efficient while promoting goals that policymakers have identified as important, such as increasing college enrollment or homeownership.

(3) The Tax Treatment of Health Care: Shifting to a Progressive Tax Credit

One of the most important roles that tax policy plays is in health care. Health spending is 17 percent of the economy. The tax treatment of health care plays an important role in how the roughly half of this spending that is private, $1.3 trillion in 2008, is spent. It also affects the number of Americans that are uninsured, currently estimated at 47 million by the Census Bureau. In fact, the quantitative magnitude of the inefficiencies associated with the tax treatment of healthcare may exceed the inefficiencies associated with many of the other traditional concerns of tax policy, like a broader base and flatter rates.

Healthcare is subsidized through the tax code in a number of ways, but the most important of these is that employer contributions to insurance premiums are excluded from earnings for the purposes of determining income and payroll taxes. The federal cost of this exclusion and other tax benefits for healthcare is about $200 billion annually, roughly the same as federal spending on Medicaid. This tax treatment, originally a historical accident, creates several non-neutralities in the tax code:

• More favoritism for purchasing health insurance than other goods. The most important non-neutrality introduced by the tax treatment is an incentive to purchase health insurance rather than other goods. In effect the tax treatment reduces the after-tax price of health insurance for the typical worker by about 20 percent. In the absence of this nonneutrality the number of Americans without insurance would likely be substantially larger.

• More favoritism for health insurance through employers than through the individual market. If your employer does not offer health insurance, or you choose not to buy it, you cannot get the same tax advantages when you purchase insurance in the individual market. This, together with other historical accidents and objective advantages of employer-sponsored insurance, underpins the employer-sponsored system that provides insurance to 177 million, or 88 percent, of the privately insured. This non-neutrality in the tax system helps pool people together, solving the adverse selection problem and other market failures that would otherwise impair the health insurance market. But there is a serious question as to whether the tax non-neutrality is the most efficient or desirable way to solve these problems.

• More of a subsidy for insurance for high-income households than low-income households. As discussed above, a tax deduction or exclusion provides a larger subsidy for households in higher tax brackets. For example, consider an employer contributing $8,000 to a family policy and requiring a $2,000 contribution by the worker. A low-income worker facing a marginal tax rate of 10 percent, the typical marginal tax rate for a worker at the 30^th percentile, would have to give up $9,200 in after-tax income for this policy -- effectively an 8 percent subsidy for insurance.⁵ In contrast, a high-income family might be in the 40 percent marginal rate -- and thus have to give up $6,800 in after-tax income for this policy -- effectively a 32 percent subsidy for insurance. If the lower-income worker receives a smaller tax-advantaged employer contribution to his or her insurance the disparity will be even greater. This is not only unfair, but it also leads firms with disproportionate numbers of lower-income workers to be less likely to offer insurance or pay a large fraction of the premium, both of which lead to less insurance coverage for lower-income workers.⁶ If the goal of the tax subsidy is to increase the number of Americans with insurance then this form of provision is inefficient because the current subsidy is evidently too small to encourage low-income people to demand insurance and is likely higher than it needs to be to ensure that high-income people are covered.

• More of an incentive to spend money on healthcare. The exclusion and other tax benefits for health care reduce the after-tax cost of that spending, leading to more spending on health care and less spending on everything else than would be the case in the absence of these incentives. The design of the current tax incentive magnifies this effect because the combination of the employer exclusion with the general lack of a tax deduction for out-of-pocket expenses leads to insurance plans with lower co-payments and deductibles and thus higher spending. Two studies suggest that eliminating the tax exclusion for health insurance premiums could result in a 41 to 65 percent increase in the coinsurance rate, which could lead to anywhere from a 9 to 38 percent reduction in health expenditures for the privately insured. Both economic theory and evidence from the RAND health insurance experiment and other studies suggest that such a reduction in spending would result in little if any worsening in health outcomes.

(4) Discouraging Undesired Activity: the Role of Pigouvian Taxes

Just as it can sometimes be appropriate to introduce non-neutralities into the tax system to encourage desired activities so to can it be appropriate to use the tax system to discourage undesirable ones like the smoking, drinking alcohol, or emitting carbon. In this manner, so-called Pigouvian taxes can lead businesses and consumers to take the social costs of their actions into account, helping to ensure that the outcome of decentralized decisions and market competition leads to overall social efficiency. Today, for example, gasoline is taxed at both the federal and the state level, but the evidence is that these taxes fall short of neutralizing the external harm associated with gasoline consumption, which includes not only climate change but also congestion, traffic accidents, and increased economic vulnerability to supply disruptions. Meanwhile the production of electricity and other energy from coal and natural gas is not taxed at all, despite its large contribution to climate change.

The climate problem could be addressed directly through the tax code by implementing a carbon tax that is combined with other tax cuts to ensure that it is revenue- and distribution-neutral, thus protecting low- and moderate-income families who would otherwise have a hard time paying higher energy bills.⁷ Alternatively, the government could issue a limited number of permits for emitting carbon and allow them to be traded, a so-called "cap-and-trade system."⁸ Like a carbon tax, a cap-and-trade system would effectively put a price on carbon that would be passed on to consumers -- making it desirable to combine it with an auction system for the permits and a lump-sum compensation mechanism for households.

Another motivation for non-neutral taxes is when a myopic individual takes insufficient account of the harm that immediate actions have on his or her long-term well-being, a notion that economist Jonathan Gruber has termed an "internality."⁹ In this case, the tax reduces consumption and has potentially large benefits for the individuals involved -- benefits that well exceed the cost of the taxes.

(5) Corporate and Capital Taxes

Lately significant attention has been focused on the fact that the United States now has the second highest corporate tax rate in the world. Less attention has been focused on the fact that the United States simultaneously has, in recent years, averaged the fourth lowest corporate tax collections as a share of the economy of any OECD country. According to the Treasury:

The debate over the optimal tax rate on capital income is highly contentious. According to the Congressional Budget Office (CBO) the overall effective tax rate on capital income is 14 percent.¹¹ Proponents of a pure income tax believe that capital income should be taxed at the same rate of labor income, which is somewhat higher than this rate. They justify this argument by noting that total income is the best measure of ability to pay and that having different tax rates on different forms of income encourages sheltering and other avoidance activity.¹² In contrast, proponents of a consumption tax argue that the tax on capital income should be set at zero to avoid discouraging savings and investment.

But regardless of one's stance on the question of whether the 14 percent rate is too low or too high, there is no justification for the highly variable tax rates on different forms of capital income shown in Table 1. If a corporation finances its investment by borrowing it can take advantage of accelerated depreciation of its investments, deduct its interest, and pay a substantial fraction of the interest and dividends on its proceeds to tax indifferent parties like retirement accounts and foundations. The result is that the tax rate on debt-financed corporate investment is -6 percent. In other words, the tax system is subsidizing debt-financed corporate investment. In contrast, equity-financed corporate investment is taxed at the corporate and individual level and faces a combined rate of 36 percent.

 Table 1. Effective Tax Rates on Capital Income

The tax system is also highly non-neutral towards other forms of capital investment. Noncorporate investment is, on average, taxed more lightly than corporate investment, discouraging the use of the corporate form. Housing is heavily tax favored, and owner-occupied housing faces a negative tax rate. This can lead to the over-building of and over-borrowing on houses.

A complex and inconsistent set of depreciation rules means that different assets are taxed at very different rates ranging from above 30 percent for computers and manufacturing structures to 10 percent or lower for mining, petroleum and natural gas structures. This encourages underinvestment in some areas and overinvestment in other areas, reducing the long-run productivity of the U.S. economy.

Finally, the corporate tax system is not neutral relative to investment in the United States and investment abroad. In this case there are a number of competing and sometimes contradictory neutrality concepts.¹³ One of the classic concepts is "capital export neutrality." The current system violates this form of neutrality because it gives a tax advantage to overseas investment in the form of a deferral that is not available to domestic investment. Specifically, the U.S. government taxes U.S. multinationals on income earned both at home and abroad. Income earned abroad, however, is generally not taxed until it is repatriated (and firms receive a credit for any foreign taxes paid on that income). This "deferral" of taxation allows foreign-earned income to grow tax free, distorts investment decisions, potentially leads to overinvestment abroad, creates an incentive for firms to earn (or report) profits in low tax countries, and reduces U.S. corporate tax revenue. Making the tax code more neutral with respect to international transactions could also be done in a manner that lowers the corporate tax rate overall.¹⁴

Moving towards more neutral taxation of business income need not require increasing the deficit or reducing the overall progressivity of the tax systems. There are a number of models policymakers could contribute, the most comprehensive being the Business Enterprise Income Tax (BEIT) developed by Edward Kleinbard.¹⁵

Conclusion

The longstanding shortcomings in the tax code are compounded by three major imminent issues: the 2010 expiration date for the tax cuts enacted since 2001, the expansion of the Alternative Minimum Tax, and the worldwide trend towards lower corporate tax rates. These issues give policymakers an opportunity to reform the tax code.

A number of considerations are important in such a reform. The large increase in income inequality provides a rationale for making the tax code more progressive. The tax code could be substantially simpler. And, the topic of my testimony today, the tax code could be more efficient if it were more neutral vis-à-vis different economic activities and if deviations from neutrality were better designed.

² Public Law 93-344, The Congressional Budget Act of 1974.

³ Note that these totals are indicative of the extent of tax expenditures but are not an estimate of the revenue that would be raised by repealing these tax expenditures because they ignore behavioral effects and the interaction of tax expenditures with other provisions in the tax code and other tax expenditures.

⁴ Lily L. Batchelder, Fred T. Goldberg, Jr., and Peter R. Orszag. 2006. "Efficiency and Tax incentives: The Case for Refundable Tax Credits." 59 Stanford Law Review 23.

⁵ These marginal tax rates are consistent with the ones estimated by Congressional Budget Office, "Effective Marginal Tax Rates on Labor Income," (2005). The CBO estimates are adjusted to reflect the additional Social Security benefits accrued as a result of having a higher taxable income. This is both the economically correct treatment and is also consistent with the budgetary treatment of Social Security used in estimating the $200 billion cost of the tax expenditures for healthcare.

⁶ This tax incentive is compounded by the availability of Medicaid and the State Children's Health Insurance Program (SCHIP) to lower-income workers or their dependents.

⁷ Gilbert E. Metcalf. 2007. "A Proposal for a U.S. Carbon Tax Swap: An Equitable Tax Reform to Address Global Climate Change." Discussion Paper 2007-12, The Hamilton Project, Washington, DC.

⁸ Robert N. Stavins. 2007. "A U.S. Cap-and-Trade System to Address Global Climate Change." Discussion Paper 2007-13, The Hamilton Project, Washington, DC.

⁹ Jonathan Gruber. 2002. "The Economics of Tobacco Regulation." Health Affairs, Vol 21(2).

¹⁰ U.S. Department of the Treasury. 2007. "Approaches to Improve the Competitiveness of the U.S. Business Tax System for the 21st Century." Washington, DC.

¹¹ Congressional Budget Office. 2005. "Taxing Capital Income: Effective Rates and Approaches to Reform." Washington, DC.

¹² A more recent economic literature also justifies a positive tax on capital income as a way to proxy for the underlying ability of the individual or to provide insurance against wage shocks, although in these cases the tax would not necessarily be equal to the tax on labor income. See Emanuel Saez. 2002. "The Desirability of Commodity Taxation under Non-Linear Income Taxation and Heterogeneous Tastes." Journal of Public Economics 83: 217 -- 230. See also Shinichi Nishiyama and Kent Smetters. 2005. "Consumption Taxes and Economic Efficiency with Idiosyncratic Wage Shocks." Journal of Political Economic, Vol 113(5). And Mikhail Golosov, Aleh Tsyvinski, and Ivan Werning. 2006. "New Dynamic Public Finance: A User's Guide." NBER Macroeconomics Annual.

¹³ Mihir A. Desai, and James R. Hines Jr. 2003. "Evaluating International Tax Reform." National Tax Journal, Vol 56(3).

¹⁴ Rosanne Altschuler and Harry Grubert. 2006. "Corporate Taxes in the World Economy: Reforming the Taxation of Cross-Border Income."

¹⁵ Edward D. Kleinbard. 2007. "Rehabilitating the Business Income Tax." Discussion Paper 2007-09, The Hamilton Project, Washington, DC.